Skip to contentSkip to conversation

server side user xattr support (RFC 8276)

PATCH In review Frank van der Linden 6 participants 28 messages Jun 23, 2020 – Jul 29, 2020

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

v3 cover server side user xattr support (RFC 8276)

v3:
* Rebase to v5.8-rc2
* Use length probe + allocate + query for the listxattr and setxattr operations to avoid allocating unneeded space.
* Because of the above, drop the 'use kvmalloc for svcxdr_tmpalloc' patch, as it's no longer needed.

v2:
* As per the discussion, user extended attributes are enabled if the client and server support them (e.g. they support 4.2 and advertise the user extended attribute FATTR). There are no longer options to switch them off.
* The code is no longer conditioned on a config option.
* The number of patches has been reduced somewhat by merging smaller, related ones.
* Renamed some functions and added parameter comments as requested.

v1:

* Split in to client and server (changed from the original RFC patch).

Original RFC combined set is here:

https://www.spinics.net/lists/linux-nfs/msg74843.html

In general, these patches were, both server and client, tested as follows:
* stress-ng-xattr with 1000 workers
* Test all corner cases (XATTR_SIZE_*)
* Test all failure cases (no xattr, setxattr with different or invalid flags, etc).
* Verify the content of xattrs across several operations.
* Use KASAN and KMEMLEAK for a longer mix of testruns to verify that there were no leaks (after unmounting the filesystem).
* Interop run against FreeBSD server/client implementation.
* Ran xfstests-dev, with no unexpected/new failures as compared to an unpatched kernel. To fully use xfstests-dev, it needed some modifications, as it expects to either use all xattr namespaces, or none. Whereas NFS only suppors the "user." namespace (+ optional ACLs). I will send the changes in seperately.

Frank van der Linden (10):
xattr: break delegations in {set,remove}xattr
xattr: add a function to check if a namespace is supported nfs,nfsd: NFSv4.2 extended attribute protocol definitions
nfsd: split off the write decode code in to a separate function
nfsd: add defines for NFSv4.2 extended attribute support
nfsd: define xattr functions to call in to their vfs counterparts
nfsd: take xattr bits in to account for permission checks
nfsd: add structure definitions for xattr requests / responses
nfsd: implement the xattr functions and en/decode logic
nfsd: add fattr support for user extended attributes

 fs/nfsd/nfs4proc.c        | 128 ++++++++-
 fs/nfsd/nfs4xdr.c         | 531 +++++++++++++++++++++++++++++++++++---
 fs/nfsd/nfsd.h            |   5 +-
 fs/nfsd/vfs.c             | 239 +++++++++++++++++
 fs/nfsd/vfs.h             |  10 +
 fs/nfsd/xdr4.h            |  31 +++
 fs/xattr.c                | 111 +++++++-
 include/linux/nfs4.h      |  22 +-
 include/linux/xattr.h     |   4 +
 include/uapi/linux/nfs4.h |   3 +
 10 files changed, 1044 insertions(+), 40 deletions(-)
Signature · 1 line
2.17.2
#1 lore raw

PATCH v3 1/10 xattr: break delegations in {set,remove}xattr

set/removexattr on an exported filesystem should break NFS delegations. This is true in general, but also for the upcoming support for RFC 8726 (NFSv4 extended attribute support). Make sure that they do.

Additonally, they need to grow a _locked variant, since callers might call this with i_rwsem held (like the NFS server code).

---
 fs/xattr.c            | 84 +++++++++++++++++++++++++++++++++++++++----
 include/linux/xattr.h |  2 ++
 2 files changed, 79 insertions(+), 7 deletions(-)
fs/xattr.c +77 -7
index 91608d9bfc6a..95f38f57347f 100644
@@ -204,10 +204,22 @@ int __vfs_setxattr_noperm(struct dentry *dentry, const char *name,
 	return error;
 }
 
-
+/**
+ * __vfs_setxattr_locked: set an extended attribute while holding the inode
+ * lock
+ *
+ *  @dentry - object to perform setxattr on
+ *  @name - xattr name to set
+ *  @value - value to set @name to
+ *  @size - size of @value
+ *  @flags - flags to pass into filesystem operations
+ *  @delegated_inode - on return, will contain an inode pointer that
+ *  a delegation was broken on, NULL if none.
+ */
 int
-vfs_setxattr(struct dentry *dentry, const char *name, const void *value,
-		size_t size, int flags)
+__vfs_setxattr_locked(struct dentry *dentry, const char *name,
+		const void *value, size_t size, int flags,
+		struct inode **delegated_inode)
 {
 	struct inode *inode = dentry->d_inode;
 	int error;
@@ -216,15 +228,40 @@ vfs_setxattr(struct dentry *dentry, const char *name, const void *value,
 	if (error)
 		return error;
 
-	inode_lock(inode);
 	error = security_inode_setxattr(dentry, name, value, size, flags);
 	if (error)
 		goto out;
 
+	error = try_break_deleg(inode, delegated_inode);
+	if (error)
+		goto out;
+
 	error = __vfs_setxattr_noperm(dentry, name, value, size, flags);
 
 out:
+	return error;
+}
+EXPORT_SYMBOL_GPL(__vfs_setxattr_locked);
+
+int
+vfs_setxattr(struct dentry *dentry, const char *name, const void *value,
+		size_t size, int flags)
+{
+	struct inode *inode = dentry->d_inode;
+	struct inode *delegated_inode = NULL;
+	int error;
+
+retry_deleg:
+	inode_lock(inode);
+	error = __vfs_setxattr_locked(dentry, name, value, size, flags,
+	    &delegated_inode);
 	inode_unlock(inode);
+
+	if (delegated_inode) {
+		error = break_deleg_wait(&delegated_inode);
+		if (!error)
+			goto retry_deleg;
+	}
 	return error;
 }
 EXPORT_SYMBOL_GPL(vfs_setxattr);
@@ -378,8 +415,18 @@ __vfs_removexattr(struct dentry *dentry, const char *name)
 }
 EXPORT_SYMBOL(__vfs_removexattr);
 
+/**
+ * __vfs_removexattr_locked: set an extended attribute while holding the inode
+ * lock
+ *
+ *  @dentry - object to perform setxattr on
+ *  @name - name of xattr to remove
+ *  @delegated_inode - on return, will contain an inode pointer that
+ *  a delegation was broken on, NULL if none.
+ */
 int
-vfs_removexattr(struct dentry *dentry, const char *name)
+__vfs_removexattr_locked(struct dentry *dentry, const char *name,
+		struct inode **delegated_inode)
 {
 	struct inode *inode = dentry->d_inode;
 	int error;
@@ -388,11 +435,14 @@ vfs_removexattr(struct dentry *dentry, const char *name)
 	if (error)
 		return error;
 
-	inode_lock(inode);
 	error = security_inode_removexattr(dentry, name);
 	if (error)
 		goto out;
 
+	error = try_break_deleg(inode, delegated_inode);
+	if (error)
+		goto out;
+
 	error = __vfs_removexattr(dentry, name);
 
 	if (!error) {
@@ -401,12 +451,32 @@ vfs_removexattr(struct dentry *dentry, const char *name)
 	}
 
 out:
+	return error;
+}
+EXPORT_SYMBOL_GPL(__vfs_removexattr_locked);
+
+int
+vfs_removexattr(struct dentry *dentry, const char *name)
+{
+	struct inode *inode = dentry->d_inode;
+	struct inode *delegated_inode = NULL;
+	int error;
+
+retry_deleg:
+	inode_lock(inode);
+	error = __vfs_removexattr_locked(dentry, name, &delegated_inode);
 	inode_unlock(inode);
+
+	if (delegated_inode) {
+		error = break_deleg_wait(&delegated_inode);
+		if (!error)
+			goto retry_deleg;
+	}
+
 	return error;
 }
 EXPORT_SYMBOL_GPL(vfs_removexattr);
 
-
 /*
  * Extended attribute SET operations
  */
include/linux/xattr.h +2 -0
index 47eaa34f8761..a2f3cd02653c 100644
@@ -51,8 +51,10 @@ ssize_t vfs_getxattr(struct dentry *, const char *, void *, size_t);
 ssize_t vfs_listxattr(struct dentry *d, char *list, size_t size);
 int __vfs_setxattr(struct dentry *, struct inode *, const char *, const void *, size_t, int);
 int __vfs_setxattr_noperm(struct dentry *, const char *, const void *, size_t, int);
+int __vfs_setxattr_locked(struct dentry *, const char *, const void *, size_t, int, struct inode **);
 int vfs_setxattr(struct dentry *, const char *, const void *, size_t, int);
 int __vfs_removexattr(struct dentry *, const char *);
+int __vfs_removexattr_locked(struct dentry *, const char *, struct inode **);
 int vfs_removexattr(struct dentry *, const char *);
 
 ssize_t generic_listxattr(struct dentry *dentry, char *buffer, size_t buffer_size);
Signature · 1 line
2.17.2
#2 lore raw

Re: [PATCH v3 01/10] xattr: break delegations in {set,remove}xattr

Hi Al,

Do you have any comments / concerns about this patch? It's part of nfs server side user xattr support, full series here:

https://lore.kernel.org/linux-nfs/[email protected]/

I copied this one to linux-fsdevel and you, just giving you an extra ping. Bruce/Chuck are OK with the rest of the series, so I just need your ACK on this one, and the next one (will send the ping separately).

Thanks,

- Frank

Quoted 176 lines · Frank van der Linden

set/removexattr on an exported filesystem should break NFS delegations. This is true in general, but also for the upcoming support for RFC 8726 (NFSv4 extended attribute support). Make sure that they do.

Additonally, they need to grow a _locked variant, since callers might call this with i_rwsem held (like the NFS server code).

---
 fs/xattr.c            | 84 +++++++++++++++++++++++++++++++++++++++----
 include/linux/xattr.h |  2 ++
 2 files changed, 79 insertions(+), 7 deletions(-)
fs/xattr.c +77 -7
index 91608d9bfc6a..95f38f57347f 100644
@@ -204,10 +204,22 @@ int __vfs_setxattr_noperm(struct dentry *dentry, const char *name,
 	return error;
 }
 
-
+/**
+ * __vfs_setxattr_locked: set an extended attribute while holding the inode
+ * lock
+ *
+ *  @dentry - object to perform setxattr on
+ *  @name - xattr name to set
+ *  @value - value to set @name to
+ *  @size - size of @value
+ *  @flags - flags to pass into filesystem operations
+ *  @delegated_inode - on return, will contain an inode pointer that
+ *  a delegation was broken on, NULL if none.
+ */
 int
-vfs_setxattr(struct dentry *dentry, const char *name, const void *value,
-		size_t size, int flags)
+__vfs_setxattr_locked(struct dentry *dentry, const char *name,
+		const void *value, size_t size, int flags,
+		struct inode **delegated_inode)
 {
 	struct inode *inode = dentry->d_inode;
 	int error;
@@ -216,15 +228,40 @@ vfs_setxattr(struct dentry *dentry, const char *name, const void *value,
 	if (error)
 		return error;
 
-	inode_lock(inode);
 	error = security_inode_setxattr(dentry, name, value, size, flags);
 	if (error)
 		goto out;
 
+	error = try_break_deleg(inode, delegated_inode);
+	if (error)
+		goto out;
+
 	error = __vfs_setxattr_noperm(dentry, name, value, size, flags);
 
 out:
+	return error;
+}
+EXPORT_SYMBOL_GPL(__vfs_setxattr_locked);
+
+int
+vfs_setxattr(struct dentry *dentry, const char *name, const void *value,
+		size_t size, int flags)
+{
+	struct inode *inode = dentry->d_inode;
+	struct inode *delegated_inode = NULL;
+	int error;
+
+retry_deleg:
+	inode_lock(inode);
+	error = __vfs_setxattr_locked(dentry, name, value, size, flags,
+	    &delegated_inode);
 	inode_unlock(inode);
+
+	if (delegated_inode) {
+		error = break_deleg_wait(&delegated_inode);
+		if (!error)
+			goto retry_deleg;
+	}
 	return error;
 }
 EXPORT_SYMBOL_GPL(vfs_setxattr);
@@ -378,8 +415,18 @@ __vfs_removexattr(struct dentry *dentry, const char *name)
 }
 EXPORT_SYMBOL(__vfs_removexattr);
 
+/**
+ * __vfs_removexattr_locked: set an extended attribute while holding the inode
+ * lock
+ *
+ *  @dentry - object to perform setxattr on
+ *  @name - name of xattr to remove
+ *  @delegated_inode - on return, will contain an inode pointer that
+ *  a delegation was broken on, NULL if none.
+ */
 int
-vfs_removexattr(struct dentry *dentry, const char *name)
+__vfs_removexattr_locked(struct dentry *dentry, const char *name,
+		struct inode **delegated_inode)
 {
 	struct inode *inode = dentry->d_inode;
 	int error;
@@ -388,11 +435,14 @@ vfs_removexattr(struct dentry *dentry, const char *name)
 	if (error)
 		return error;
 
-	inode_lock(inode);
 	error = security_inode_removexattr(dentry, name);
 	if (error)
 		goto out;
 
+	error = try_break_deleg(inode, delegated_inode);
+	if (error)
+		goto out;
+
 	error = __vfs_removexattr(dentry, name);
 
 	if (!error) {
@@ -401,12 +451,32 @@ vfs_removexattr(struct dentry *dentry, const char *name)
 	}
 
 out:
+	return error;
+}
+EXPORT_SYMBOL_GPL(__vfs_removexattr_locked);
+
+int
+vfs_removexattr(struct dentry *dentry, const char *name)
+{
+	struct inode *inode = dentry->d_inode;
+	struct inode *delegated_inode = NULL;
+	int error;
+
+retry_deleg:
+	inode_lock(inode);
+	error = __vfs_removexattr_locked(dentry, name, &delegated_inode);
 	inode_unlock(inode);
+
+	if (delegated_inode) {
+		error = break_deleg_wait(&delegated_inode);
+		if (!error)
+			goto retry_deleg;
+	}
+
 	return error;
 }
 EXPORT_SYMBOL_GPL(vfs_removexattr);
 
-
 /*
  * Extended attribute SET operations
  */
include/linux/xattr.h +2 -0
index 47eaa34f8761..a2f3cd02653c 100644
@@ -51,8 +51,10 @@ ssize_t vfs_getxattr(struct dentry *, const char *, void *, size_t);
 ssize_t vfs_listxattr(struct dentry *d, char *list, size_t size);
 int __vfs_setxattr(struct dentry *, struct inode *, const char *, const void *, size_t, int);
 int __vfs_setxattr_noperm(struct dentry *, const char *, const void *, size_t, int);
+int __vfs_setxattr_locked(struct dentry *, const char *, const void *, size_t, int, struct inode **);
 int vfs_setxattr(struct dentry *, const char *, const void *, size_t, int);
 int __vfs_removexattr(struct dentry *, const char *);
+int __vfs_removexattr_locked(struct dentry *, const char *, struct inode **);
 int vfs_removexattr(struct dentry *, const char *);
 
 ssize_t generic_listxattr(struct dentry *dentry, char *buffer, size_t buffer_size);
Signature · 1 line
2.17.2
#3 lore raw

Re: [PATCH v3 01/10] xattr: break delegations in {set,remove}xattr

Quoted 7 lines · Frank van der Linden

Hi Al,

Do you have any comments / concerns about this patch? It's part of nfs server side user xattr support, full series here:

https://lore.kernel.org/linux-nfs/[email protected]/

Quoted 7 lines

I copied this one to linux-fsdevel and you, just giving you an extra ping. Bruce/Chuck are OK with the rest of the series, so I just need your ACK on this one, and the next one (will send the ping separately).

Thanks,

- Frank

Hi Al,

Any comments on this one?

It's a simple change to break (NFSv4) delegations on set/removexattr, so it's pretty nfsd specific.

Linus - since this is nfsd specific, could this go in through the nfsd maintainer tree directly? Chuck has included it in a tree that is being readied for 5.9.

Thanks,

- Frank

Quoted 179 lines · Frank van der Linden
Quoted 176 lines

set/removexattr on an exported filesystem should break NFS delegations. This is true in general, but also for the upcoming support for RFC 8726 (NFSv4 extended attribute support). Make sure that they do.

Additonally, they need to grow a _locked variant, since callers might call this with i_rwsem held (like the NFS server code).

---
 fs/xattr.c            | 84 +++++++++++++++++++++++++++++++++++++++----
 include/linux/xattr.h |  2 ++
 2 files changed, 79 insertions(+), 7 deletions(-)
fs/xattr.c +77 -7
index 91608d9bfc6a..95f38f57347f 100644
@@ -204,10 +204,22 @@ int __vfs_setxattr_noperm(struct dentry *dentry, const char *name,
 	return error;
 }
 
-
+/**
+ * __vfs_setxattr_locked: set an extended attribute while holding the inode
+ * lock
+ *
+ *  @dentry - object to perform setxattr on
+ *  @name - xattr name to set
+ *  @value - value to set @name to
+ *  @size - size of @value
+ *  @flags - flags to pass into filesystem operations
+ *  @delegated_inode - on return, will contain an inode pointer that
+ *  a delegation was broken on, NULL if none.
+ */
 int
-vfs_setxattr(struct dentry *dentry, const char *name, const void *value,
-		size_t size, int flags)
+__vfs_setxattr_locked(struct dentry *dentry, const char *name,
+		const void *value, size_t size, int flags,
+		struct inode **delegated_inode)
 {
 	struct inode *inode = dentry->d_inode;
 	int error;
@@ -216,15 +228,40 @@ vfs_setxattr(struct dentry *dentry, const char *name, const void *value,
 	if (error)
 		return error;
 
-	inode_lock(inode);
 	error = security_inode_setxattr(dentry, name, value, size, flags);
 	if (error)
 		goto out;
 
+	error = try_break_deleg(inode, delegated_inode);
+	if (error)
+		goto out;
+
 	error = __vfs_setxattr_noperm(dentry, name, value, size, flags);
 
 out:
+	return error;
+}
+EXPORT_SYMBOL_GPL(__vfs_setxattr_locked);
+
+int
+vfs_setxattr(struct dentry *dentry, const char *name, const void *value,
+		size_t size, int flags)
+{
+	struct inode *inode = dentry->d_inode;
+	struct inode *delegated_inode = NULL;
+	int error;
+
+retry_deleg:
+	inode_lock(inode);
+	error = __vfs_setxattr_locked(dentry, name, value, size, flags,
+	    &delegated_inode);
 	inode_unlock(inode);
+
+	if (delegated_inode) {
+		error = break_deleg_wait(&delegated_inode);
+		if (!error)
+			goto retry_deleg;
+	}
 	return error;
 }
 EXPORT_SYMBOL_GPL(vfs_setxattr);
@@ -378,8 +415,18 @@ __vfs_removexattr(struct dentry *dentry, const char *name)
 }
 EXPORT_SYMBOL(__vfs_removexattr);
 
+/**
+ * __vfs_removexattr_locked: set an extended attribute while holding the inode
+ * lock
+ *
+ *  @dentry - object to perform setxattr on
+ *  @name - name of xattr to remove
+ *  @delegated_inode - on return, will contain an inode pointer that
+ *  a delegation was broken on, NULL if none.
+ */
 int
-vfs_removexattr(struct dentry *dentry, const char *name)
+__vfs_removexattr_locked(struct dentry *dentry, const char *name,
+		struct inode **delegated_inode)
 {
 	struct inode *inode = dentry->d_inode;
 	int error;
@@ -388,11 +435,14 @@ vfs_removexattr(struct dentry *dentry, const char *name)
 	if (error)
 		return error;
 
-	inode_lock(inode);
 	error = security_inode_removexattr(dentry, name);
 	if (error)
 		goto out;
 
+	error = try_break_deleg(inode, delegated_inode);
+	if (error)
+		goto out;
+
 	error = __vfs_removexattr(dentry, name);
 
 	if (!error) {
@@ -401,12 +451,32 @@ vfs_removexattr(struct dentry *dentry, const char *name)
 	}
 
 out:
+	return error;
+}
+EXPORT_SYMBOL_GPL(__vfs_removexattr_locked);
+
+int
+vfs_removexattr(struct dentry *dentry, const char *name)
+{
+	struct inode *inode = dentry->d_inode;
+	struct inode *delegated_inode = NULL;
+	int error;
+
+retry_deleg:
+	inode_lock(inode);
+	error = __vfs_removexattr_locked(dentry, name, &delegated_inode);
 	inode_unlock(inode);
+
+	if (delegated_inode) {
+		error = break_deleg_wait(&delegated_inode);
+		if (!error)
+			goto retry_deleg;
+	}
+
 	return error;
 }
 EXPORT_SYMBOL_GPL(vfs_removexattr);
 
-
 /*
  * Extended attribute SET operations
  */
include/linux/xattr.h +2 -0
index 47eaa34f8761..a2f3cd02653c 100644
@@ -51,8 +51,10 @@ ssize_t vfs_getxattr(struct dentry *, const char *, void *, size_t);
 ssize_t vfs_listxattr(struct dentry *d, char *list, size_t size);
 int __vfs_setxattr(struct dentry *, struct inode *, const char *, const void *, size_t, int);
 int __vfs_setxattr_noperm(struct dentry *, const char *, const void *, size_t, int);
+int __vfs_setxattr_locked(struct dentry *, const char *, const void *, size_t, int, struct inode **);
 int vfs_setxattr(struct dentry *, const char *, const void *, size_t, int);
 int __vfs_removexattr(struct dentry *, const char *);
+int __vfs_removexattr_locked(struct dentry *, const char *, struct inode **);
 int vfs_removexattr(struct dentry *, const char *);
 
 ssize_t generic_listxattr(struct dentry *dentry, char *buffer, size_t buffer_size);
Signature · 1 line
2.17.2
maintainer
#4 lore raw

Re: [PATCH v3 01/10] xattr: break delegations in {set,remove}xattr

In reply to Frank van der Linden

Hi

[This is an automated email]

This commit has been processed because it contains a -stable tag. The stable tag indicates that it's relevant for the following trees: all

The bot has tested the following trees: v5.7.6, v5.4.49, v4.19.130, v4.14.186, v4.9.228, v4.4.228.

v5.7.6: Build OK!
v5.4.49: Build OK!
v4.19.130: Build OK!
v4.14.186: Build OK!
v4.9.228: Build OK!
v4.4.228: Failed to apply! Possible dependencies:

    5d6c31910bc07 ("xattr: Add __vfs_{get,set,remove}xattr helpers")
    6b2553918d8b4 ("replace ->follow_link() with new method that could stay in RCU mode")
    aa80deab33a8f ("namei: page_getlink() and page_follow_link_light() are the same thing")
    cd3417c8fc950 ("kill free_page_put_link()")
    ce23e64013348 ("->getxattr(): pass dentry and inode as separate arguments")
    fceef393a5381 ("switch ->get_link() to delayed_call, kill ->put_link()")

NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

Signature · 2 lines
Thanks
Sasha
maintainer
#5 lore raw

Re: [PATCH v3 01/10] xattr: break delegations in {set,remove}xattr

In reply to Frank van der Linden

Hi

[This is an automated email]

This commit has been processed because it contains a -stable tag. The stable tag indicates that it's relevant for the following trees: all

The bot has tested the following trees: v5.7.6, v5.4.49, v4.19.130, v4.14.186, v4.9.228, v4.4.228.

v5.7.6: Build OK!
v5.4.49: Build OK!
v4.19.130: Build OK!
v4.14.186: Build OK!
v4.9.228: Build OK!
v4.4.228: Failed to apply! Possible dependencies:

    5d6c31910bc07 ("xattr: Add __vfs_{get,set,remove}xattr helpers")
    6b2553918d8b4 ("replace ->follow_link() with new method that could stay in RCU mode")
    aa80deab33a8f ("namei: page_getlink() and page_follow_link_light() are the same thing")
    cd3417c8fc950 ("kill free_page_put_link()")
    ce23e64013348 ("->getxattr(): pass dentry and inode as separate arguments")
    fceef393a5381 ("switch ->get_link() to delayed_call, kill ->put_link()")

NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

Signature · 2 lines
Thanks
Sasha
#6 lore raw

Re: [PATCH v3 01/10] xattr: break delegations in {set,remove}xattr

Hi Sasha-

Quoted 25 lines · Sasha Levin

Hi

[This is an automated email]

This commit has been processed because it contains a -stable tag. The stable tag indicates that it's relevant for the following trees: all

The bot has tested the following trees: v5.7.6, v5.4.49, v4.19.130, v4.14.186, v4.9.228, v4.4.228.

v5.7.6: Build OK!
v5.4.49: Build OK!
v4.19.130: Build OK!
v4.14.186: Build OK!
v4.9.228: Build OK!
v4.4.228: Failed to apply! Possible dependencies:
5d6c31910bc07 ("xattr: Add __vfs_{get,set,remove}xattr helpers") 6b2553918d8b4 ("replace ->follow_link() with new method that could stay in RCU mode") aa80deab33a8f ("namei: page_getlink() and page_follow_link_light() are the same thing") cd3417c8fc950 ("kill free_page_put_link()")
ce23e64013348 ("->getxattr(): pass dentry and inode as separate arguments") fceef393a5381 ("switch ->get_link() to delayed_call, kill ->put_link()")

Quoted 3 lines

NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

I've updated the "cc: stable" tag in my testing branch to include "# v4.9+".

--
Chuck Lever

#7 lore raw

PATCH v3 2/10 xattr: add a function to check if a namespace is supported

In reply to Frank van der Linden

Add a function that checks is an extended attribute namespace is supported for an inode, meaning that a handler must be present for either the whole namespace, or at least one synthetic xattr in the namespace.

To be used by the nfs server code when being queried for extended attributes support.

---
 fs/xattr.c            | 27 +++++++++++++++++++++++++++
 include/linux/xattr.h |  2 ++
 2 files changed, 29 insertions(+)
fs/xattr.c +27 -0
index 95f38f57347f..386b45676d7e 100644
@@ -134,6 +134,33 @@ xattr_permission(struct inode *inode, const char *name, int mask)
 	return inode_permission(inode, mask);
 }
 
+/*
+ * Look for any handler that deals with the specified namespace.
+ */
+int
+xattr_supported_namespace(struct inode *inode, const char *prefix)
+{
+	const struct xattr_handler **handlers = inode->i_sb->s_xattr;
+	const struct xattr_handler *handler;
+	size_t preflen;
+
+	if (!(inode->i_opflags & IOP_XATTR)) {
+		if (unlikely(is_bad_inode(inode)))
+			return -EIO;
+		return -EOPNOTSUPP;
+	}
+
+	preflen = strlen(prefix);
+
+	for_each_xattr_handler(handlers, handler) {
+		if (!strncmp(xattr_prefix(handler), prefix, preflen))
+			return 0;
+	}
+
+	return -EOPNOTSUPP;
+}
+EXPORT_SYMBOL(xattr_supported_namespace);
+
 int
 __vfs_setxattr(struct dentry *dentry, struct inode *inode, const char *name,
 	       const void *value, size_t size, int flags)
include/linux/xattr.h +2 -0
index a2f3cd02653c..fac75810d9d3 100644
@@ -61,6 +61,8 @@ ssize_t generic_listxattr(struct dentry *dentry, char *buffer, size_t buffer_siz
 ssize_t vfs_getxattr_alloc(struct dentry *dentry, const char *name,
 			   char **xattr_value, size_t size, gfp_t flags);
 
+int xattr_supported_namespace(struct inode *inode, const char *prefix);
+
 static inline const char *xattr_prefix(const struct xattr_handler *handler)
 {
 	return handler->prefix ?: handler->name;
Signature · 1 line
2.17.2
#8 lore raw

Re: [PATCH v3 02/10] xattr: add a function to check if a namespace is supported

Hi Al,

..and here is the other xattr change that was part of the nfsd user xattr support series that I mentioned in my previous email.

Full series here:

https://lore.kernel.org/linux-nfs/[email protected]/

Any comments / concerned about this one?

Thanks,

- Frank

Quoted 70 lines · Frank van der Linden

Add a function that checks is an extended attribute namespace is supported for an inode, meaning that a handler must be present for either the whole namespace, or at least one synthetic xattr in the namespace.

To be used by the nfs server code when being queried for extended attributes support.

---
 fs/xattr.c            | 27 +++++++++++++++++++++++++++
 include/linux/xattr.h |  2 ++
 2 files changed, 29 insertions(+)
fs/xattr.c +27 -0
index 95f38f57347f..386b45676d7e 100644
@@ -134,6 +134,33 @@ xattr_permission(struct inode *inode, const char *name, int mask)
 	return inode_permission(inode, mask);
 }
 
+/*
+ * Look for any handler that deals with the specified namespace.
+ */
+int
+xattr_supported_namespace(struct inode *inode, const char *prefix)
+{
+	const struct xattr_handler **handlers = inode->i_sb->s_xattr;
+	const struct xattr_handler *handler;
+	size_t preflen;
+
+	if (!(inode->i_opflags & IOP_XATTR)) {
+		if (unlikely(is_bad_inode(inode)))
+			return -EIO;
+		return -EOPNOTSUPP;
+	}
+
+	preflen = strlen(prefix);
+
+	for_each_xattr_handler(handlers, handler) {
+		if (!strncmp(xattr_prefix(handler), prefix, preflen))
+			return 0;
+	}
+
+	return -EOPNOTSUPP;
+}
+EXPORT_SYMBOL(xattr_supported_namespace);
+
 int
 __vfs_setxattr(struct dentry *dentry, struct inode *inode, const char *name,
 	       const void *value, size_t size, int flags)
include/linux/xattr.h +2 -0
index a2f3cd02653c..fac75810d9d3 100644
@@ -61,6 +61,8 @@ ssize_t generic_listxattr(struct dentry *dentry, char *buffer, size_t buffer_siz
 ssize_t vfs_getxattr_alloc(struct dentry *dentry, const char *name,
 			   char **xattr_value, size_t size, gfp_t flags);
 
+int xattr_supported_namespace(struct inode *inode, const char *prefix);
+
 static inline const char *xattr_prefix(const struct xattr_handler *handler)
 {
 	return handler->prefix ?: handler->name;
Signature · 1 line
2.17.2
#9 lore raw

Re: [PATCH v3 02/10] xattr: add a function to check if a namespace is supported

Quoted 7 lines · Frank van der Linden

Hi Al,

..and here is the other xattr change that was part of the nfsd user xattr support series that I mentioned in my previous email.

Full series here:

Quoted 7 lines

https://lore.kernel.org/linux-nfs/[email protected]/

Any comments / concerned about this one?

Thanks,

- Frank

Hi Al,

Here's the other one I'm just sending a quick ping on. It's a simple change - just add a little new function that enables nfsd to check if the "user." namespace is at all supported by a filesystem.

Any comments?

Again, Linus - this is a pretty small change, doesn't affect any existing codepaths, and it's already in the tree Chuck is setting up for 5.9. Could this go in through that directly?

Thanks,

- Frank

Quoted 72 lines · Frank van der Linden
Quoted 70 lines

Add a function that checks is an extended attribute namespace is supported for an inode, meaning that a handler must be present for either the whole namespace, or at least one synthetic xattr in the namespace.

To be used by the nfs server code when being queried for extended attributes support.

---
 fs/xattr.c            | 27 +++++++++++++++++++++++++++
 include/linux/xattr.h |  2 ++
 2 files changed, 29 insertions(+)
fs/xattr.c +27 -0
index 95f38f57347f..386b45676d7e 100644
@@ -134,6 +134,33 @@ xattr_permission(struct inode *inode, const char *name, int mask)
 	return inode_permission(inode, mask);
 }
 
+/*
+ * Look for any handler that deals with the specified namespace.
+ */
+int
+xattr_supported_namespace(struct inode *inode, const char *prefix)
+{
+	const struct xattr_handler **handlers = inode->i_sb->s_xattr;
+	const struct xattr_handler *handler;
+	size_t preflen;
+
+	if (!(inode->i_opflags & IOP_XATTR)) {
+		if (unlikely(is_bad_inode(inode)))
+			return -EIO;
+		return -EOPNOTSUPP;
+	}
+
+	preflen = strlen(prefix);
+
+	for_each_xattr_handler(handlers, handler) {
+		if (!strncmp(xattr_prefix(handler), prefix, preflen))
+			return 0;
+	}
+
+	return -EOPNOTSUPP;
+}
+EXPORT_SYMBOL(xattr_supported_namespace);
+
 int
 __vfs_setxattr(struct dentry *dentry, struct inode *inode, const char *name,
 	       const void *value, size_t size, int flags)
include/linux/xattr.h +2 -0
index a2f3cd02653c..fac75810d9d3 100644
@@ -61,6 +61,8 @@ ssize_t generic_listxattr(struct dentry *dentry, char *buffer, size_t buffer_siz
 ssize_t vfs_getxattr_alloc(struct dentry *dentry, const char *name,
 			   char **xattr_value, size_t size, gfp_t flags);
 
+int xattr_supported_namespace(struct inode *inode, const char *prefix);
+
 static inline const char *xattr_prefix(const struct xattr_handler *handler)
 {
 	return handler->prefix ?: handler->name;
Signature · 1 line
2.17.2
#10 lore raw

Re: [PATCH v3 02/10] xattr: add a function to check if a namespace is supported

Quoted 4 lines · Frank van der Linden

Again, Linus - this is a pretty small change, doesn't affect any existing codepaths, and it's already in the tree Chuck is setting up for 5.9. Could this go in through that directly?

Both ok by me, but I'd like to have Al ack them. Al?

Linus

#11 lore raw

Re: [PATCH v3 02/10] xattr: add a function to check if a namespace is supported

Hi Linus-

Quoted 10 lines · Linus Torvalds
Quoted 4 lines · Frank van der Linden

Again, Linus - this is a pretty small change, doesn't affect any existing codepaths, and it's already in the tree Chuck is setting up for 5.9. Could this go in through that directly?

Both ok by me, but I'd like to have Al ack them. Al?

I have the NFSD user xattr patches in the current series waiting to be merged into v5.9. I'd like to create the nfsd-5.9 merge tag soon, but I haven't heard any review comments from Al. How would you like me to proceed?

--
Chuck Lever

maintainer
#12 lore raw

Re: [PATCH v3 02/10] xattr: add a function to check if a namespace is supported

Quoted 11 lines · Chuck Lever

Hi Linus-

Quoted 9 lines · Linus Torvalds
Quoted 4 lines · Frank van der Linden

Again, Linus - this is a pretty small change, doesn't affect any existing codepaths, and it's already in the tree Chuck is setting up for 5.9. Could this go in through that directly?

Quoted 6 lines
Quoted 1 line

Both ok by me, but I'd like to have Al ack them. Al?

I have the NFSD user xattr patches in the current series waiting to be merged into v5.9. I'd like to create the nfsd-5.9 merge tag soon, but I haven't heard any review comments from Al. How would you like me to proceed?

Looks sane, AFAICS. Sorry, still digging myself from under the mounds of mail...

#13 lore raw

Re: [PATCH v3 02/10] xattr: add a function to check if a namespace is supported

Quoted 18 lines · Al Viro
Quoted 13 lines · Chuck Lever

Hi Linus-

Quoted 10 lines · Linus Torvalds
Quoted 4 lines · Frank van der Linden

Again, Linus - this is a pretty small change, doesn't affect any existing codepaths, and it's already in the tree Chuck is setting up for 5.9. Could this go in through that directly?

Both ok by me, but I'd like to have Al ack them. Al?

Quoted 7 lines
Quoted 4 lines

I have the NFSD user xattr patches in the current series waiting to be merged into v5.9. I'd like to create the nfsd-5.9 merge tag soon, but I haven't heard any review comments from Al. How would you like me to proceed?

Looks sane, AFAICS. Sorry, still digging myself from under the mounds of mail...

No problem at all. Thanks!

--
Chuck Lever

#14 lore raw

PATCH v3 3/10 nfs,nfsd: NFSv4.2 extended attribute protocol definitions

In reply to Frank van der Linden

Add definitions for the new operations, errors and flags as defined in RFC 8276 (File System Extended Attributes in NFSv4).

---
 include/linux/nfs4.h      | 20 ++++++++++++++++++++
 include/uapi/linux/nfs4.h |  3 +++
 2 files changed, 23 insertions(+)
include/linux/nfs4.h +20 -0
index 4dba3c948932..e6ca9d1d2e76 100644
@@ -150,6 +150,12 @@ enum nfs_opnum4 {
 	OP_WRITE_SAME = 70,
 	OP_CLONE = 71,
 
+	/* xattr support (RFC8726) */
+	OP_GETXATTR                = 72,
+	OP_SETXATTR                = 73,
+	OP_LISTXATTRS              = 74,
+	OP_REMOVEXATTR             = 75,
+
 	OP_ILLEGAL = 10044,
 };
 
@@ -280,6 +286,10 @@ enum nfsstat4 {
 	NFS4ERR_WRONG_LFS = 10092,
 	NFS4ERR_BADLABEL = 10093,
 	NFS4ERR_OFFLOAD_NO_REQS = 10094,
+
+	/* xattr (RFC8276) */
+	NFS4ERR_NOXATTR        = 10095,
+	NFS4ERR_XATTR2BIG      = 10096,
 };
 
 static inline bool seqid_mutating_err(u32 err)
@@ -452,6 +462,7 @@ enum change_attr_type4 {
 #define FATTR4_WORD2_CHANGE_ATTR_TYPE	(1UL << 15)
 #define FATTR4_WORD2_SECURITY_LABEL     (1UL << 16)
 #define FATTR4_WORD2_MODE_UMASK		(1UL << 17)
+#define FATTR4_WORD2_XATTR_SUPPORT	(1UL << 18)
 
 /* MDS threshold bitmap bits */
 #define THRESHOLD_RD                    (1UL << 0)
@@ -700,4 +711,13 @@ struct nl4_server {
 		struct nfs42_netaddr	nl4_addr; /* NL4_NETADDR */
 	} u;
 };
+
+/*
+ * Options for setxattr. These match the flags for setxattr(2).
+ */
+enum nfs4_setxattr_options {
+	SETXATTR4_EITHER	= 0,
+	SETXATTR4_CREATE	= 1,
+	SETXATTR4_REPLACE	= 2,
+};
 #endif
include/uapi/linux/nfs4.h +3 -0
index 8572930cf5b0..bf197e99b98f 100644
@@ -33,6 +33,9 @@
 #define NFS4_ACCESS_EXTEND      0x0008
 #define NFS4_ACCESS_DELETE      0x0010
 #define NFS4_ACCESS_EXECUTE     0x0020
+#define NFS4_ACCESS_XAREAD      0x0040
+#define NFS4_ACCESS_XAWRITE     0x0080
+#define NFS4_ACCESS_XALIST      0x0100
 
 #define NFS4_FH_PERSISTENT		0x0000
 #define NFS4_FH_NOEXPIRE_WITH_OPEN	0x0001
Signature · 1 line
2.17.2
#15 lore raw

PATCH v3 4/10 nfsd: split off the write decode code in to a separate function

In reply to Frank van der Linden

nfs4_decode_write has code to parse incoming XDR write data in to a kvec head, and a list of pages.

Put this code in to a separate function, so that it can be used later by the xattr code, for setxattr. No functional change.

---
 fs/nfsd/nfs4xdr.c | 72 +++++++++++++++++++++++++++--------------------
 1 file changed, 42 insertions(+), 30 deletions(-)
fs/nfsd/nfs4xdr.c +42 -30
index 996ac01ee977..48806b493eba 100644
@@ -257,6 +257,44 @@ svcxdr_dupstr(struct nfsd4_compoundargs *argp, void *buf, u32 len)
 	return p;
 }
 
+static __be32
+svcxdr_construct_vector(struct nfsd4_compoundargs *argp, struct kvec *head,
+			struct page ***pagelist, u32 buflen)
+{
+	int avail;
+	int len;
+	int pages;
+
+	/* Sorry .. no magic macros for this.. *
+	 * READ_BUF(write->wr_buflen);
+	 * SAVEMEM(write->wr_buf, write->wr_buflen);
+	 */
+	avail = (char *)argp->end - (char *)argp->p;
+	if (avail + argp->pagelen < buflen) {
+		dprintk("NFSD: xdr error (%s:%d)\n",
+			       __FILE__, __LINE__);
+		return nfserr_bad_xdr;
+	}
+	head->iov_base = argp->p;
+	head->iov_len = avail;
+	*pagelist = argp->pagelist;
+
+	len = XDR_QUADLEN(buflen) << 2;
+	if (len >= avail) {
+		len -= avail;
+
+		pages = len >> PAGE_SHIFT;
+		argp->pagelist += pages;
+		argp->pagelen -= pages * PAGE_SIZE;
+		len -= pages * PAGE_SIZE;
+
+		next_decode_page(argp);
+	}
+	argp->p += XDR_QUADLEN(len);
+
+	return 0;
+}
+
 /**
  * savemem - duplicate a chunk of memory for later processing
  * @argp: NFSv4 compound argument structure to be freed with
@@ -1265,8 +1303,6 @@ nfsd4_decode_verify(struct nfsd4_compoundargs *argp, struct nfsd4_verify *verify
 static __be32
 nfsd4_decode_write(struct nfsd4_compoundargs *argp, struct nfsd4_write *write)
 {
-	int avail;
-	int len;
 	DECODE_HEAD;
 
 	status = nfsd4_decode_stateid(argp, &write->wr_stateid);
@@ -1279,34 +1315,10 @@ nfsd4_decode_write(struct nfsd4_compoundargs *argp, struct nfsd4_write *write)
 		goto xdr_error;
 	write->wr_buflen = be32_to_cpup(p++);
 
-	/* Sorry .. no magic macros for this.. *
-	 * READ_BUF(write->wr_buflen);
-	 * SAVEMEM(write->wr_buf, write->wr_buflen);
-	 */
-	avail = (char*)argp->end - (char*)argp->p;
-	if (avail + argp->pagelen < write->wr_buflen) {
-		dprintk("NFSD: xdr error (%s:%d)\n",
-				__FILE__, __LINE__);
-		goto xdr_error;
-	}
-	write->wr_head.iov_base = p;
-	write->wr_head.iov_len = avail;
-	write->wr_pagelist = argp->pagelist;
-
-	len = XDR_QUADLEN(write->wr_buflen) << 2;
-	if (len >= avail) {
-		int pages;
-
-		len -= avail;
-
-		pages = len >> PAGE_SHIFT;
-		argp->pagelist += pages;
-		argp->pagelen -= pages * PAGE_SIZE;
-		len -= pages * PAGE_SIZE;
-
-		next_decode_page(argp);
-	}
-	argp->p += XDR_QUADLEN(len);
+	status = svcxdr_construct_vector(argp, &write->wr_head,
+					 &write->wr_pagelist, write->wr_buflen);
+	if (status)
+		return status;
 
 	DECODE_TAIL;
 }
Signature · 1 line
2.17.2
#16 lore raw

PATCH v3 5/10 nfsd: add defines for NFSv4.2 extended attribute support

In reply to Frank van der Linden

Add defines for server-side extended attribute support. Most have already been added as part of client support, but these are the network order error codes for the noxattr and xattr2big errors, and the addition of the xattr support to the supported file attributes (if configured).

---
 fs/nfsd/nfsd.h | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
fs/nfsd/nfsd.h +4 -1
index 36cdd81b6688..5343c771da18 100644
@@ -283,6 +283,8 @@ void		nfsd_lockd_shutdown(void);
 #define nfserr_wrong_lfs		cpu_to_be32(NFS4ERR_WRONG_LFS)
 #define nfserr_badlabel			cpu_to_be32(NFS4ERR_BADLABEL)
 #define nfserr_file_open		cpu_to_be32(NFS4ERR_FILE_OPEN)
+#define nfserr_xattr2big		cpu_to_be32(NFS4ERR_XATTR2BIG)
+#define nfserr_noxattr			cpu_to_be32(NFS4ERR_NOXATTR)
 
 /* error codes for internal use */
 /* if a request fails due to kmalloc failure, it gets dropped.
@@ -384,7 +386,8 @@ void		nfsd_lockd_shutdown(void);
 	(NFSD4_1_SUPPORTED_ATTRS_WORD2 | \
 	FATTR4_WORD2_CHANGE_ATTR_TYPE | \
 	FATTR4_WORD2_MODE_UMASK | \
-	NFSD4_2_SECURITY_ATTRS)
+	NFSD4_2_SECURITY_ATTRS | \
+	FATTR4_WORD2_XATTR_SUPPORT)
 
 extern const u32 nfsd_suppattrs[3][3];
 
Signature · 1 line
2.17.2
#17 lore raw

PATCH v3 6/10 nfsd: define xattr functions to call in to their vfs counterparts

In reply to Frank van der Linden

This adds the filehandle based functions for the xattr operations that call in to the vfs layer to do the actual work.

---
 fs/nfsd/vfs.c | 227 ++++++++++++++++++++++++++++++++++++++++++++++++++
 fs/nfsd/vfs.h |  10 +++
 2 files changed, 237 insertions(+)
fs/nfsd/vfs.c +227 -0
index c3fbab1753ec..b1dd8690e25d 100644
@@ -2059,6 +2059,233 @@ static int exp_rdonly(struct svc_rqst *rqstp, struct svc_export *exp)
 	return nfsexp_flags(rqstp, exp) & NFSEXP_READONLY;
 }
 
+#ifdef CONFIG_NFSD_V4
+/*
+ * Helper function to translate error numbers. In the case of xattr operations,
+ * some error codes need to be translated outside of the standard translations.
+ *
+ * ENODATA needs to be translated to nfserr_noxattr.
+ * E2BIG to nfserr_xattr2big.
+ *
+ * Additionally, vfs_listxattr can return -ERANGE. This means that the
+ * file has too many extended attributes to retrieve inside an
+ * XATTR_LIST_MAX sized buffer. This is a bug in the xattr implementation:
+ * filesystems will allow the adding of extended attributes until they hit
+ * their own internal limit. This limit may be larger than XATTR_LIST_MAX.
+ * So, at that point, the attributes are present and valid, but can't
+ * be retrieved using listxattr, since the upper level xattr code enforces
+ * the XATTR_LIST_MAX limit.
+ *
+ * This bug means that we need to deal with listxattr returning -ERANGE. The
+ * best mapping is to return TOOSMALL.
+ */
+static __be32
+nfsd_xattr_errno(int err)
+{
+	switch (err) {
+	case -ENODATA:
+		return nfserr_noxattr;
+	case -E2BIG:
+		return nfserr_xattr2big;
+	case -ERANGE:
+		return nfserr_toosmall;
+	}
+	return nfserrno(err);
+}
+
+/*
+ * Retrieve the specified user extended attribute. To avoid always
+ * having to allocate the maximum size (since we are not getting
+ * a maximum size from the RPC), do a probe + alloc. Hold a reader
+ * lock on i_rwsem to prevent the extended attribute from changing
+ * size while we're doing this.
+ */
+__be32
+nfsd_getxattr(struct svc_rqst *rqstp, struct svc_fh *fhp, char *name,
+	      void **bufp, int *lenp)
+{
+	ssize_t len;
+	__be32 err;
+	char *buf;
+	struct inode *inode;
+	struct dentry *dentry;
+
+	err = fh_verify(rqstp, fhp, 0, NFSD_MAY_READ);
+	if (err)
+		return err;
+
+	err = nfs_ok;
+	dentry = fhp->fh_dentry;
+	inode = d_inode(dentry);
+
+	inode_lock_shared(inode);
+
+	len = vfs_getxattr(dentry, name, NULL, 0);
+
+	/*
+	 * Zero-length attribute, just return.
+	 */
+	if (len == 0) {
+		*bufp = NULL;
+		*lenp = 0;
+		goto out;
+	}
+
+	if (len < 0) {
+		err = nfsd_xattr_errno(len);
+		goto out;
+	}
+
+	if (len > *lenp) {
+		err = nfserr_toosmall;
+		goto out;
+	}
+
+	buf = kvmalloc(len, GFP_KERNEL | GFP_NOFS);
+	if (buf == NULL) {
+		err = nfserr_jukebox;
+		goto out;
+	}
+
+	len = vfs_getxattr(dentry, name, buf, len);
+	if (len <= 0) {
+		kvfree(buf);
+		buf = NULL;
+		err = nfsd_xattr_errno(len);
+	}
+
+	*lenp = len;
+	*bufp = buf;
+
+out:
+	inode_unlock_shared(inode);
+
+	return err;
+}
+
+/*
+ * Retrieve the xattr names. Since we can't know how many are
+ * user extended attributes, we must get all attributes here,
+ * and have the XDR encode filter out the "user." ones.
+ *
+ * While this could always just allocate an XATTR_LIST_MAX
+ * buffer, that's a waste, so do a probe + allocate. To
+ * avoid any changes between the probe and allocate, wrap
+ * this in inode_lock.
+ */
+__be32
+nfsd_listxattr(struct svc_rqst *rqstp, struct svc_fh *fhp, char **bufp,
+	       int *lenp)
+{
+	ssize_t len;
+	__be32 err;
+	char *buf;
+	struct inode *inode;
+	struct dentry *dentry;
+
+	err = fh_verify(rqstp, fhp, 0, NFSD_MAY_READ);
+	if (err)
+		return err;
+
+	dentry = fhp->fh_dentry;
+	inode = d_inode(dentry);
+	*lenp = 0;
+
+	inode_lock_shared(inode);
+
+	len = vfs_listxattr(dentry, NULL, 0);
+	if (len <= 0) {
+		err = nfsd_xattr_errno(len);
+		goto out;
+	}
+
+	if (len > XATTR_LIST_MAX) {
+		err = nfserr_xattr2big;
+		goto out;
+	}
+
+	/*
+	 * We're holding i_rwsem - use GFP_NOFS.
+	 */
+	buf = kvmalloc(len, GFP_KERNEL | GFP_NOFS);
+	if (buf == NULL) {
+		err= nfserr_jukebox;
+		goto out;
+	}
+
+	len = vfs_listxattr(dentry, buf, len);
+	if (len <= 0) {
+		kvfree(buf);
+		err = nfsd_xattr_errno(len);
+		goto out;
+	}
+
+	*lenp = len;
+	*bufp = buf;
+
+	err = nfs_ok;
+out:
+	inode_unlock_shared(inode);
+
+	return err;
+}
+
+/*
+ * Removexattr and setxattr need to call fh_lock to both lock the inode
+ * and set the change attribute. Since the top-level vfs_removexattr
+ * and vfs_setxattr calls already do their own inode_lock calls, call
+ * the _locked variant. Pass in a NULL pointer for delegated_inode,
+ * and let the client deal with NFS4ERR_DELAY (same as with e.g.
+ * setattr and remove).
+ */
+__be32
+nfsd_removexattr(struct svc_rqst *rqstp, struct svc_fh *fhp, char *name)
+{
+	int err, ret;
+
+	err = fh_verify(rqstp, fhp, 0, NFSD_MAY_WRITE);
+	if (err)
+		return err;
+
+	ret = fh_want_write(fhp);
+	if (ret)
+		return nfserrno(ret);
+
+	fh_lock(fhp);
+
+	ret = __vfs_removexattr_locked(fhp->fh_dentry, name, NULL);
+
+	fh_unlock(fhp);
+	fh_drop_write(fhp);
+
+	return nfsd_xattr_errno(ret);
+}
+
+__be32
+nfsd_setxattr(struct svc_rqst *rqstp, struct svc_fh *fhp, char *name,
+	      void *buf, u32 len, u32 flags)
+{
+	int err, ret;
+
+	err = fh_verify(rqstp, fhp, 0, NFSD_MAY_WRITE);
+	if (err)
+		return err;
+
+	ret = fh_want_write(fhp);
+	if (ret)
+		return nfserrno(ret);
+	fh_lock(fhp);
+
+	ret = __vfs_setxattr_locked(fhp->fh_dentry, name, buf, len, flags,
+				    NULL);
+
+	fh_unlock(fhp);
+	fh_drop_write(fhp);
+
+	return nfsd_xattr_errno(ret);
+}
+#endif
+
 /*
  * Check for a user's access permissions to this inode.
  */
fs/nfsd/vfs.h +10 -0
index 3eb660ad80d1..a2442ebe5acf 100644
@@ -76,6 +76,16 @@ __be32		do_nfsd_create(struct svc_rqst *, struct svc_fh *,
 __be32		nfsd_commit(struct svc_rqst *, struct svc_fh *,
 				loff_t, unsigned long, __be32 *verf);
 #endif /* CONFIG_NFSD_V3 */
+#ifdef CONFIG_NFSD_V4
+__be32		nfsd_getxattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
+			    char *name, void **bufp, int *lenp);
+__be32		nfsd_listxattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
+			    char **bufp, int *lenp);
+__be32		nfsd_removexattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
+			    char *name);
+__be32		nfsd_setxattr(struct svc_rqst *rqstp, struct svc_fh *fhp,
+			    char *name, void *buf, u32 len, u32 flags);
+#endif
 int 		nfsd_open_break_lease(struct inode *, int);
 __be32		nfsd_open(struct svc_rqst *, struct svc_fh *, umode_t,
 				int, struct file **);
Signature · 1 line
2.17.2
#18 lore raw

PATCH v3 7/10 nfsd: take xattr bits in to account for permission checks

In reply to Frank van der Linden

Since the NFSv4.2 extended attributes extension defines 3 new access bits for xattr operations, take them in to account when validating what the client is asking for, and when checking permissions.

---
 fs/nfsd/nfs4proc.c |  8 +++++++-
 fs/nfsd/vfs.c      | 12 ++++++++++++
 2 files changed, 19 insertions(+), 1 deletion(-)
fs/nfsd/nfs4proc.c +7 -1
index a09c35f0f6f0..841aad772798 100644
@@ -566,8 +566,14 @@ nfsd4_access(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	     union nfsd4_op_u *u)
 {
 	struct nfsd4_access *access = &u->access;
+	u32 access_full;
 
-	if (access->ac_req_access & ~NFS3_ACCESS_FULL)
+	access_full = NFS3_ACCESS_FULL;
+	if (cstate->minorversion >= 2)
+		access_full |= NFS4_ACCESS_XALIST | NFS4_ACCESS_XAREAD |
+			       NFS4_ACCESS_XAWRITE;
+
+	if (access->ac_req_access & ~access_full)
 		return nfserr_inval;
 
 	access->ac_resp_access = access->ac_req_access;
fs/nfsd/vfs.c +12 -0
index b1dd8690e25d..8e223b3bf26f 100644
@@ -612,6 +612,12 @@ static struct accessmap	nfs3_regaccess[] = {
     {	NFS3_ACCESS_MODIFY,	NFSD_MAY_WRITE|NFSD_MAY_TRUNC	},
     {	NFS3_ACCESS_EXTEND,	NFSD_MAY_WRITE			},
 
+#ifdef CONFIG_NFSD_V4
+    {	NFS4_ACCESS_XAREAD,	NFSD_MAY_READ			},
+    {	NFS4_ACCESS_XAWRITE,	NFSD_MAY_WRITE			},
+    {	NFS4_ACCESS_XALIST,	NFSD_MAY_READ			},
+#endif
+
     {	0,			0				}
 };
 
@@ -622,6 +628,12 @@ static struct accessmap	nfs3_diraccess[] = {
     {	NFS3_ACCESS_EXTEND,	NFSD_MAY_EXEC|NFSD_MAY_WRITE	},
     {	NFS3_ACCESS_DELETE,	NFSD_MAY_REMOVE			},
 
+#ifdef CONFIG_NFSD_V4
+    {	NFS4_ACCESS_XAREAD,	NFSD_MAY_READ			},
+    {	NFS4_ACCESS_XAWRITE,	NFSD_MAY_WRITE			},
+    {	NFS4_ACCESS_XALIST,	NFSD_MAY_READ			},
+#endif
+
     {	0,			0				}
 };
 
Signature · 1 line
2.17.2
#19 lore raw

PATCH v3 8/10 nfsd: add structure definitions for xattr requests / responses

In reply to Frank van der Linden

Add the structures used in extended attribute request / response handling.

---
 fs/nfsd/xdr4.h | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)
fs/nfsd/xdr4.h +31 -0
index db63d39b1507..66499fb6b567 100644
@@ -224,6 +224,32 @@ struct nfsd4_putfh {
 	bool		no_verify;	    /* represents foreigh fh */
 };
 
+struct nfsd4_getxattr {
+	char		*getxa_name;		/* request */
+	u32		getxa_len;		/* request */
+	void		*getxa_buf;
+};
+
+struct nfsd4_setxattr {
+	u32		setxa_flags;		/* request */
+	char		*setxa_name;		/* request */
+	char		*setxa_buf;		/* request */
+	u32		setxa_len;		/* request */
+	struct nfsd4_change_info  setxa_cinfo;	/* response */
+};
+
+struct nfsd4_removexattr {
+	char		*rmxa_name;		/* request */
+	struct nfsd4_change_info  rmxa_cinfo;	/* response */
+};
+
+struct nfsd4_listxattrs {
+	u64		lsxa_cookie;		/* request */
+	u32		lsxa_maxcount;		/* request */
+	char		*lsxa_buf;		/* unfiltered buffer (reply) */
+	u32		lsxa_len;		/* unfiltered len (reply) */
+};
+
 struct nfsd4_open {
 	u32		op_claim_type;      /* request */
 	struct xdr_netobj op_fname;	    /* request - everything but CLAIM_PREV */
@@ -649,6 +675,11 @@ struct nfsd4_op {
 		struct nfsd4_offload_status	offload_status;
 		struct nfsd4_copy_notify	copy_notify;
 		struct nfsd4_seek		seek;
+
+		struct nfsd4_getxattr		getxattr;
+		struct nfsd4_setxattr		setxattr;
+		struct nfsd4_listxattrs		listxattrs;
+		struct nfsd4_removexattr	removexattr;
 	} u;
 	struct nfs4_replay *			replay;
 };
Signature · 1 line
2.17.2
#20 lore raw

PATCH v3 9/10 nfsd: implement the xattr functions and en/decode logic

In reply to Frank van der Linden

Implement the main entry points for the *XATTR operations.

Add functions to calculate the reply size for the user extended attribute operations, and implement the XDR encode / decode logic for these operations.

Add the user extended attributes operations to nfsd4_ops.

---
 fs/nfsd/nfs4proc.c   | 120 ++++++++++++
 fs/nfsd/nfs4xdr.c    | 450 +++++++++++++++++++++++++++++++++++++++++++
 include/linux/nfs4.h |   2 +-
 3 files changed, 571 insertions(+), 1 deletion(-)
fs/nfsd/nfs4proc.c +120 -0
index 841aad772798..a527da3d8052 100644
@@ -2097,6 +2097,68 @@ nfsd4_layoutreturn(struct svc_rqst *rqstp,
 }
 #endif /* CONFIG_NFSD_PNFS */
 
+static __be32
+nfsd4_getxattr(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
+	       union nfsd4_op_u *u)
+{
+	struct nfsd4_getxattr *getxattr = &u->getxattr;
+
+	return nfsd_getxattr(rqstp, &cstate->current_fh,
+			     getxattr->getxa_name, &getxattr->getxa_buf,
+			     &getxattr->getxa_len);
+}
+
+static __be32
+nfsd4_setxattr(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
+	   union nfsd4_op_u *u)
+{
+	struct nfsd4_setxattr *setxattr = &u->setxattr;
+	__be32 ret;
+
+	if (opens_in_grace(SVC_NET(rqstp)))
+		return nfserr_grace;
+
+	ret = nfsd_setxattr(rqstp, &cstate->current_fh, setxattr->setxa_name,
+			    setxattr->setxa_buf, setxattr->setxa_len,
+			    setxattr->setxa_flags);
+
+	if (!ret)
+		set_change_info(&setxattr->setxa_cinfo, &cstate->current_fh);
+
+	return ret;
+}
+
+static __be32
+nfsd4_listxattrs(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
+	   union nfsd4_op_u *u)
+{
+	/*
+	 * Get the entire list, then copy out only the user attributes
+	 * in the encode function.
+	 */
+	return nfsd_listxattr(rqstp, &cstate->current_fh,
+			     &u->listxattrs.lsxa_buf, &u->listxattrs.lsxa_len);
+}
+
+static __be32
+nfsd4_removexattr(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
+	   union nfsd4_op_u *u)
+{
+	struct nfsd4_removexattr *removexattr = &u->removexattr;
+	__be32 ret;
+
+	if (opens_in_grace(SVC_NET(rqstp)))
+		return nfserr_grace;
+
+	ret = nfsd_removexattr(rqstp, &cstate->current_fh,
+	    removexattr->rmxa_name);
+
+	if (!ret)
+		set_change_info(&removexattr->rmxa_cinfo, &cstate->current_fh);
+
+	return ret;
+}
+
 /*
  * NULL call.
  */
@@ -2706,6 +2768,42 @@ static inline u32 nfsd4_seek_rsize(struct svc_rqst *rqstp, struct nfsd4_op *op)
 	return (op_encode_hdr_size + 3) * sizeof(__be32);
 }
 
+static inline u32 nfsd4_getxattr_rsize(struct svc_rqst *rqstp,
+				       struct nfsd4_op *op)
+{
+	u32 maxcount, rlen;
+
+	maxcount = svc_max_payload(rqstp);
+	rlen = min_t(u32, XATTR_SIZE_MAX, maxcount);
+
+	return (op_encode_hdr_size + 1 + XDR_QUADLEN(rlen)) * sizeof(__be32);
+}
+
+static inline u32 nfsd4_setxattr_rsize(struct svc_rqst *rqstp,
+				       struct nfsd4_op *op)
+{
+	return (op_encode_hdr_size + op_encode_change_info_maxsz)
+		* sizeof(__be32);
+}
+static inline u32 nfsd4_listxattrs_rsize(struct svc_rqst *rqstp,
+					 struct nfsd4_op *op)
+{
+	u32 maxcount, rlen;
+
+	maxcount = svc_max_payload(rqstp);
+	rlen = min(op->u.listxattrs.lsxa_maxcount, maxcount);
+
+	return (op_encode_hdr_size + 4 + XDR_QUADLEN(rlen)) * sizeof(__be32);
+}
+
+static inline u32 nfsd4_removexattr_rsize(struct svc_rqst *rqstp,
+					  struct nfsd4_op *op)
+{
+	return (op_encode_hdr_size + op_encode_change_info_maxsz)
+		* sizeof(__be32);
+}
+
+
 static const struct nfsd4_operation nfsd4_ops[] = {
 	[OP_ACCESS] = {
 		.op_func = nfsd4_access,
@@ -3087,6 +3185,28 @@ static const struct nfsd4_operation nfsd4_ops[] = {
 		.op_name = "OP_COPY_NOTIFY",
 		.op_rsize_bop = nfsd4_copy_notify_rsize,
 	},
+	[OP_GETXATTR] = {
+		.op_func = nfsd4_getxattr,
+		.op_name = "OP_GETXATTR",
+		.op_rsize_bop = nfsd4_getxattr_rsize,
+	},
+	[OP_SETXATTR] = {
+		.op_func = nfsd4_setxattr,
+		.op_flags = OP_MODIFIES_SOMETHING | OP_CACHEME,
+		.op_name = "OP_SETXATTR",
+		.op_rsize_bop = nfsd4_setxattr_rsize,
+	},
+	[OP_LISTXATTRS] = {
+		.op_func = nfsd4_listxattrs,
+		.op_name = "OP_LISTXATTRS",
+		.op_rsize_bop = nfsd4_listxattrs_rsize,
+	},
+	[OP_REMOVEXATTR] = {
+		.op_func = nfsd4_removexattr,
+		.op_flags = OP_MODIFIES_SOMETHING | OP_CACHEME,
+		.op_name = "OP_REMOVEXATTR",
+		.op_rsize_bop = nfsd4_removexattr_rsize,
+	},
 };
 
 /**
fs/nfsd/nfs4xdr.c +450 -0

480 lines not on this page · raw message

include/linux/nfs4.h +1 -1
index e6ca9d1d2e76..33ebe476428e 100644
@@ -165,7 +165,7 @@ Needs to be updated if more operations are defined in future.*/
 #define FIRST_NFS4_OP	OP_ACCESS
 #define LAST_NFS40_OP	OP_RELEASE_LOCKOWNER
 #define LAST_NFS41_OP	OP_RECLAIM_COMPLETE
-#define LAST_NFS42_OP	OP_CLONE
+#define LAST_NFS42_OP	OP_REMOVEXATTR
 #define LAST_NFS4_OP	LAST_NFS42_OP
 
 enum nfsstat4 {
Signature · 1 line
2.17.2
#21 lore raw

PATCH v3 10/10 nfsd: add fattr support for user extended attributes

In reply to Frank van der Linden

Check if user extended attributes are supported for an inode, and return the answer when being queried for file attributes.

An exported filesystem can now signal its RFC8276 user extended attributes capability.

---
 fs/nfsd/nfs4xdr.c | 9 +++++++++
 1 file changed, 9 insertions(+)
fs/nfsd/nfs4xdr.c +9 -0
index 8bacc0ceae19..259d5ad0e3f4 100644
@@ -3213,6 +3213,15 @@ nfsd4_encode_fattr(struct xdr_stream *xdr, struct svc_fh *fhp,
 	}
 #endif
 
+	if (bmval2 & FATTR4_WORD2_XATTR_SUPPORT) {
+		p = xdr_reserve_space(xdr, 4);
+		if (!p)
+			goto out_resource;
+		err = xattr_supported_namespace(d_inode(dentry),
+						XATTR_USER_PREFIX);
+		*p++ = cpu_to_be32(err == 0);
+	}
+
 	attrlen = htonl(xdr->buf->len - attrlen_offset - 4);
 	write_bytes_to_xdr_buf(xdr->buf, attrlen_offset, &attrlen, 4);
 	status = nfs_ok;
Signature · 1 line
2.17.2
#22 lore raw
In reply to Frank van der Linden
Quoted 25 lines · Frank van der Linden

v3:
* Rebase to v5.8-rc2
* Use length probe + allocate + query for the listxattr and setxattr operations to avoid allocating unneeded space.
* Because of the above, drop the 'use kvmalloc for svcxdr_tmpalloc' patch, as it's no longer needed.

v2:
* As per the discussion, user extended attributes are enabled if the client and server support them (e.g. they support 4.2 and advertise the user extended attribute FATTR). There are no longer options to switch them off.
* The code is no longer conditioned on a config option.
* The number of patches has been reduced somewhat by merging smaller, related ones.
* Renamed some functions and added parameter comments as requested.

v1:

* Split in to client and server (changed from the original RFC patch).

Original RFC combined set is here:

https://www.spinics.net/lists/linux-nfs/msg74843.html

Quoted 7 lines

In general, these patches were, both server and client, tested as follows:
* stress-ng-xattr with 1000 workers
* Test all corner cases (XATTR_SIZE_*)
* Test all failure cases (no xattr, setxattr with different or invalid flags, etc).
* Verify the content of xattrs across several operations.

Do you have some code to share for these tests?

--b.

Quoted 35 lines
(quoted diff) +0 -0
 	* Interop run against FreeBSD server/client implementation.
 	* Ran xfstests-dev, with no unexpected/new failures as compared

to an unpatched kernel. To fully use xfstests-dev, it needed some modifications, as it expects to either use all xattr namespaces, or none. Whereas NFS only suppors the "user." namespace (+ optional ACLs). I will send the changes in seperately.

Frank van der Linden (10):
xattr: break delegations in {set,remove}xattr
xattr: add a function to check if a namespace is supported nfs,nfsd: NFSv4.2 extended attribute protocol definitions
nfsd: split off the write decode code in to a separate function
nfsd: add defines for NFSv4.2 extended attribute support
nfsd: define xattr functions to call in to their vfs counterparts
nfsd: take xattr bits in to account for permission checks
nfsd: add structure definitions for xattr requests / responses
nfsd: implement the xattr functions and en/decode logic
nfsd: add fattr support for user extended attributes

 fs/nfsd/nfs4proc.c        | 128 ++++++++-
 fs/nfsd/nfs4xdr.c         | 531 +++++++++++++++++++++++++++++++++++---
 fs/nfsd/nfsd.h            |   5 +-
 fs/nfsd/vfs.c             | 239 +++++++++++++++++
 fs/nfsd/vfs.h             |  10 +
 fs/nfsd/xdr4.h            |  31 +++
 fs/xattr.c                | 111 +++++++-
 include/linux/nfs4.h      |  22 +-
 include/linux/xattr.h     |   4 +
 include/uapi/linux/nfs4.h |   3 +
 10 files changed, 1044 insertions(+), 40 deletions(-)
Signature · 1 line
2.17.2
#23 lore raw
Quoted 11 lines · J. Bruce Fields
Quoted 7 lines

In general, these patches were, both server and client, tested as follows:
* stress-ng-xattr with 1000 workers
* Test all corner cases (XATTR_SIZE_*)
* Test all failure cases (no xattr, setxattr with different or invalid flags, etc).
* Verify the content of xattrs across several operations.

Do you have some code to share for these tests?

--b.

Sure, my main piece of test code is attached.

- Frank

#define _GNU_SOURCE
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/xattr.h>
#include <stdio.h>
#include <stdlib.h>
#include <malloc.h>
#include <unistd.h>
#include <fcntl.h>
#include <errno.h>
#include <string.h>
#include <time.h>
#include <stdarg.h>
#include <limits.h>

#ifndef XATTR_SIZE_MAX
#define XATTR_SIZE_MAX 65536
#endif

#ifndef XATTR_LIST_MAX
#define XATR_LIST_MAX 65536
#endif

#define XATTR_TEST_ATTRLEN (5 + 4 + 1) /* "user.NNNN\0" */
#define XATTR_TEST_NATTR_MAX (XATTR_LIST_MAX / XATTR_TEST_ATTRLEN)

static const char *xt_dir;

typedef struct xattr_test_handle {
	const char *xt_test;
	char xt_filename[256];
	char xt_attrkey[2 * 256];	/* Long to test limit */
	char *xt_attrbuf;
	ssize_t xt_attrlen;
	char xt_fillc;
	ssize_t xt_failoff;
	int xt_fd;
	char *xt_listbuf;
	ssize_t xt_listlen;
} xt_hdl;

#define xt_value_len(xth) ((xth)->xt_attrlen)
#define xt_value_failoff(xth) ((xth)->xt_failoff)
#define xt_fillc(xth) ((xth)->xt_fillc)
#define xt_filename(xth) ((xth)->xt_filename)

#define PAGE_SIZE sysconf(_SC_PAGE_SIZE)

static void
usage(void)

{
	fprintf(stderr, "usage: xattr <dir>\n");
	exit(1);
}

static void
fail(xt_hdl *xth, const char *fmt, ...)

{
	va_list ap;

	printf("%-24s FAIL: ", xth->xt_test);

	va_start(ap, fmt);
	vprintf(fmt, ap);
	va_end(ap);

	printf("\n");
}

static void
pass(xt_hdl *xth, const char *fmt, ...)

{
	va_list ap;

	printf("%-24s PASS: ", xth->xt_test);

	va_start(ap, fmt);
	vprintf(fmt, ap);
	va_end(ap);

	printf("\n");
}

static int
xt_set_fs(const char *dir)

{
	int ret;
	struct stat st;

	ret = stat(dir, &st);
	if (ret < 0)
		return ret;

	if (!S_ISDIR(st.st_mode))
		return -ENOTDIR;

	xt_dir = dir;

	return 0;
}

static xt_hdl *
xt_init(const char *test)

{
	xt_hdl *xth;


	if (xt_dir == NULL) {
		errno = EINVAL;
		return NULL;
	}

	xth = calloc(1, sizeof (xt_hdl));
	if (xth == NULL)
		return NULL;

	xth->xt_test = test;
	snprintf(xth->xt_filename, sizeof xth->xt_filename, "%s/%sXXXXXX",
	    xt_dir, test);

	/*
	 * Default value, can be changed.
	 */
	snprintf(xth->xt_attrkey, sizeof xth->xt_attrkey, "user.%s",
	    test);

	return xth;
}

static int
xt_file_create(xt_hdl *xth)

{
	int fd;
	int ret;

	fd = mkostemp(xth->xt_filename, O_TRUNC);
	if (fd < 0)
		return -errno;
	
	close(fd);

	return 0;
}

static void
xt_file_remove(xt_hdl *xth)

{
	unlink(xth->xt_filename);
}

static int
xt_file_open(xt_hdl *xth, int flags)

{
	int fd;

	fd = open(xth->xt_filename, flags);
	if (fd < 0)
		return errno;
	xth->xt_fd = fd;

	return fd;
}

static int
xt_file_close(xt_hdl *xth)

{
	if (xth->xt_fd != -1) {
		close(xth->xt_fd);
		xth->xt_fd = -1;
	}
}

static void
xt_value_fill(xt_hdl *xth)

{
	char c;
	char *p;
	ssize_t i;

	if (xth->xt_attrlen == 0)
		return;

	p = xth->xt_attrbuf;

	c = random() & 0xff;
	if (c == 0)
		c++;

	for (i = 0; i < xth->xt_attrlen; i++)
		p[i] = c;

	xth->xt_fillc = c;

}

static void
xt_value_free(xt_hdl *xth)

{
	if (xth->xt_attrbuf != NULL) {
		free(xth->xt_attrbuf);
		xth->xt_attrbuf = NULL;
		xth->xt_attrlen = 0;
	}
}

static int
xt_value_alloc(xt_hdl *xth, ssize_t len)

{
	char *p;

	xt_value_free(xth);

	if (len == 0)
		return 0;

	p = malloc(len);
	if (p == NULL)
		return ENOMEM;

	xth->xt_attrbuf = p;
	xth->xt_attrlen = len;
	xth->xt_failoff = 0;
	xth->xt_fd = -1;

	xt_value_fill(xth);

	return 0;
}

static int
__xt_value_check(xt_hdl *xth, ssize_t len, char c)

{
	ssize_t i;

	for (i = 0; i < len; i++) {
		if (xth->xt_attrbuf[i] != c) {
			xth->xt_failoff = i;
			return -EINVAL;
		}
	}

	return 0;
}

static int
xt_value_check(xt_hdl *xth, ssize_t len)

{
	return __xt_value_check(xth, len, xth->xt_fillc);
}

static int
xt_list_alloc(xt_hdl *xth, ssize_t size)

{
	char *p;

	p = calloc(1, size);
	if (p == NULL)
		return -ENOMEM;

	xth->xt_listbuf = p;
	xth->xt_listlen = size;

	return 0;
}

static void
xt_value_clear(xt_hdl *xth)

{
	memset(xth->xt_attrbuf, 0, xth->xt_attrlen);
}

static void
xt_list_free(xt_hdl *xth)

{
	if (xth->xt_attrbuf != NULL) {
		free(xth->xt_listbuf);
		xth->xt_listbuf = NULL;
		xth->xt_listlen = 0;
	}
}

static void
xt_free(xt_hdl *xth)

{
	xt_value_free(xth);
	xt_list_free(xth);
	free(xth);
}

static int
xt_set_attr_name(xt_hdl *xth, const char *name)

{
	if (strlen(name) > (sizeof xth->xt_attrkey) - 1)
		return -EINVAL;

	strcpy(xth->xt_attrkey, name);

	return 0;
}

static int
xt_getxattr(xt_hdl *xth)

{
	ssize_t ret;

	ret = getxattr(xth->xt_filename, xth->xt_attrkey, xth->xt_attrbuf,
	    xth->xt_attrlen);

	return ret < 0 ? -errno : ret;
}

static int
xt_fgetxattr(xt_hdl *xth)

{
	ssize_t ret;

	ret = fgetxattr(xth->xt_fd, xth->xt_attrkey, xth->xt_attrbuf,
	    xth->xt_attrlen);
	return ret < 0 ? -errno : ret;
}

static int
xt_setxattr(xt_hdl *xth, int flags)

{
	int ret;

	ret = setxattr(xth->xt_filename, xth->xt_attrkey, xth->xt_attrbuf,
	    xth->xt_attrlen, flags);

	return ret < 0 ? -errno : 0;
}

static int
xt_fsetxattr(xt_hdl *xth, int flags)

{
	int ret;

	ret = fsetxattr(xth->xt_fd, xth->xt_attrkey, xth->xt_attrbuf,
	    xth->xt_attrlen, flags);

	return ret < 0 ? -errno : 0;
}

static int
xt_removexattr(xt_hdl *xth)

{
	int ret;

	ret = removexattr(xth->xt_filename, xth->xt_attrkey);

	return ret < 0 ? -errno : 0;
}

static int
xt_fremovexattr(xt_hdl *xth)

{
	int ret;

	ret = fremovexattr(xth->xt_fd, xth->xt_attrkey);

	return ret < 0 ? -errno : 0;
}

static int
cmpstr(const void *p1, const void *p2)

{
	char *s1, *s2;

	s1 = *(char * const *) p1;
	s2 = *(char * const *) p2;

	if (s1 == s2)
		return 0;

	if (s1 == NULL)
		return 1;

	if (s2 == NULL)
		return -1;

	return strcmp(s1, s2);
}

static void
xattrs_sort(char **list, ssize_t count)

{
	qsort(list, count, sizeof (char *), cmpstr);
}

static int
xattrs_alloc(char ***listp, ssize_t count, ssize_t *lenp)

{
	ssize_t i, n, len, j;
	size_t slen;
	char **list;

	n = count;
	len = 0;
#ifdef LIST_MAX_CHECK
	if (n > (XATTR_LIST_MAX / 6))
		return -EINVAL;

#endif

	list = calloc(n, sizeof (char *));
	if (list == NULL)
		return -ENOMEM;

	for (i = 0; i < n; i++) {
#ifdef LIST_MAX_CHECK
		if ((len + 10) > XATTR_LIST_MAX)
			break;
#endif
		len += 10;

		if (asprintf(&list[i], "user.%04d", (int)i) < 0)
			break;
	}

	if (i < n) {
		for (j = 0; j < i; j++) {
			free(list[j]);
		}
		free(list);
		return -ENOMEM;
	}

	if (lenp)
		*lenp = len;
	*listp = list;

	return 0;

}

static int
xattrs_free(char **list, ssize_t count)

{
	ssize_t i;

	for (i = 0; i < count; i++) {
		if (list[i] != NULL)
			free(list[i]);
	}

	free(list);
}

static int
xattrs_buf2list(char *buf, ssize_t buflen, char ***listp, ssize_t *countp)

{
	ssize_t slen, count, len, i;
	char *p;
	char **list;

	p = buf;
	len = buflen;

	count = 0;
	while (len > 0) {
		slen = strlen(p) + 1;
		if (!strncmp(p, "user.", 5))
			count++;
		p += slen;
		len -= slen;
	}

	if (count == 0) {
		list = NULL;
		goto out;
	}

	p = buf;
	len = buflen;

	list = calloc(count, sizeof (char *));
	if (list == NULL)
		return -ENOMEM;

	i = 0;
	while (len > 0) {
		slen = strlen(p) + 1;
		if (!strncmp(p, "user.", 5))
			list[i++] = strdup(p);
		p += slen;
		len -= slen;
	}

	xattrs_sort(list, count);

out:

	*listp = list;
	*countp = count;

	return 0;
}

static int
xattrs_file2list(const char *filename, char ***listp, ssize_t *countp)

{
	ssize_t len, ret;
	int error;
	char *buf;

	len = listxattr(filename, NULL, 0);
	if (len < 0)
		return -errno;

	buf = malloc(len);
	if (buf == NULL)
		return -ENOMEM;

	ret = listxattr(filename, buf, len);
	if (ret < 0) {
		error = errno;
		*listp = NULL;
		free(buf);
		return -error;
	}

	error = xattrs_buf2list(buf, ret, listp, countp);

	free(buf);
	return error;
}

/*
* Compare two string arrays that may have NULL entries. Counts are
* the number of entries that include possible NULL entries.
*/
static int
xattrs_cmplists(char **list1, ssize_t count1, char **list2, ssize_t count2)

{
	ssize_t n1, n2;

	for (n1 = n2 = 0; n1 < count1 && n2 < count2;) {
		while (n1 < count1 && list1[n1] == NULL)
			n1++;
		while (n2 < count2 && list2[n2] == NULL)
			n2++;

		if (n1 >= count1) {
			if (n2 >= count2)
				return 0;
			return -1;
		}

		if (n2 >= count2) {
			if (n1 >= count1)
				return 0;
			return -1;
		}

		if (strcmp(list1[n1++], list2[n2++]))
			return -1;
	}

	return 0;
}

static int
xattrs_listset(const char *filename, char **list, ssize_t len)

{
	ssize_t i;
	int ret;

	for (i = 0; i < len; i++) {
		if (list[i] == NULL)
			continue;
		ret = setxattr(filename, list[i], NULL, 0, 0);
		if (ret < 0) {
			ret = -errno;
			fprintf(stderr, "%s: failed at %d (%s)\n",
					__func__, i, strerror(ret));
			return ret;
		}
	}

	return 0;
}

static int
xattrs_listrm(const char *filename, char **list, ssize_t len)

{
	ssize_t i;
	int ret;

	for (i = 0; i < len; i++) {
		if (list[i] == NULL)
			continue;
		ret = removexattr(filename, list[i]);
		if (ret < 0)
			return -errno;
	}

	return 0;
}

static void
__test_xattr_len(xt_hdl *xth, ssize_t len)

{
	int ret;
	ssize_t xlen;

	ret = xt_file_create(xth);
	if (ret < 0) {
		fail(xth, "file create errno %d (%s)", errno,
		    strerror(errno));
		return;
	}

	ret = xt_value_alloc(xth, len);
	if (ret < 0) {
		fail(xth, "value alloc errno %d (%s)", errno,
		    strerror(errno));
		goto cleanup;
	}

	ret = xt_setxattr(xth, 0);
	if (ret < 0) {
		fail(xth, "setxattr errno %d (%s)", errno, strerror(errno));
		goto cleanup;
	}

	xt_value_clear(xth);

	xlen = xt_getxattr(xth);
	if (xlen < 0) {
		fail(xth, "getxattr errno %d (%s)", errno, strerror(errno));
		goto cleanup;
	}

	if (xlen != xt_value_len(xth)) {
		fail(xth, "getxattr length mismatch (expected %lld got %lld)",
		    (long long)xt_value_len(xth), (long long)xlen);
		goto cleanup;
	}

	ret = xt_value_check(xth, len);
	if (ret < 0) {
		fail(xth, "value check failed at offset %lld",
		    xt_value_failoff(xth));
		goto cleanup;
	}

	pass(xth, "success setting / getting %lld length xattr",
	    (long long)len);
cleanup:
	xt_file_remove(xth);
}

void
test_xattr_probe()

{
	int ret;
	xt_hdl *xth;
	ssize_t xlen;
	char c;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	ret = xt_file_create(xth);
	if (ret < 0) {
		fail(xth, "file create errno %d (%s)", errno,
		    strerror(errno));
		return;
	}

	ret = xt_value_alloc(xth, 37);
	if (ret < 0) {
		fail(xth, "value alloc failure");
		goto cleanup;
	}

	c = xt_fillc(xth);

	ret = xt_setxattr(xth, 0);
	if (ret < 0) {
		fail(xth, "setxattr errno %d (%s)", errno, strerror(errno));
		goto cleanup;
	}

	ret = xt_value_alloc(xth, 0);
	if (ret < 0) {
		fail(xth, "value reset failure");
		goto cleanup;
	}

	xlen = xt_getxattr(xth);
	if (xlen < 0) {
		fail(xth, "getxattr errno %d (%s)", errno, strerror(errno));
		goto cleanup;
	}
	if (xlen != 37) {
		fail(xth, "getxattr length mismatch (expected 37, got %lld)",
		    (long long)xlen);
		goto cleanup;
	}

	ret = xt_value_alloc(xth, 37);
	if (ret < 0) {
		fail(xth, "value realloc failure");
		goto cleanup;
	}

	xlen = xt_getxattr(xth);
	if (xlen < 0) {
		fail(xth, "getxattr errno %d (%s)", errno, strerror(errno));
		goto cleanup;
	}

	if (__xt_value_check(xth, 37, c) < 0) {
		fail(xth, "value check failed at offset %lld",
		    xt_value_failoff(xth));
		goto cleanup;
	}

	pass(xth, "getxattr length probe works correctly");

cleanup:

	xt_file_remove(xth);
	xt_free(xth);
}

void
test_xattr_0()

{
	int ret;
	xt_hdl *xth;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	__test_xattr_len(xth, 0);

	xt_free(xth);
}

void
test_xattr_1()

{
	xt_hdl *xth;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	__test_xattr_len(xth, 1);

	xt_free(xth);
}

void
test_xattr_256()

{
	xt_hdl *xth;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	__test_xattr_len(xth, 256);

	xt_free(xth);
}

void
test_xattr_1page()

{
	xt_hdl *xth;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	__test_xattr_len(xth, PAGE_SIZE);

	xt_free(xth);
}

void
test_xattr_2pages()

{
	xt_hdl *xth;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	__test_xattr_len(xth, 2 * PAGE_SIZE);

	xt_free(xth);
}

void
test_xattr_max()

{
	xt_hdl *xth;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	__test_xattr_len(xth, 65536);

	xt_free(xth);
}

void
test_xattr_toolarge()

{
	xt_hdl *xth;
	ssize_t ret;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	ret = xt_file_create(xth);
	if (ret < 0) {
		fail(xth, "file create errno %d (%s)", errno,
		    strerror(errno));
		return;
	}

	ret = xt_value_alloc(xth, XATTR_SIZE_MAX + 1);
	if (ret < 0) {
		fail(xth, "value alloc errno %d (%s)", errno,
		    strerror(errno));
		goto cleanup;
	}

	ret = xt_setxattr(xth, 0);
	if (ret != -E2BIG) {
		fail(xth, "expected %d got %d (%s)", E2BIG,
		    -ret, ret == 0 ? "no error" : strerror(-ret));
		goto cleanup;
	}

	pass(xth, "setxattr beyond max length fails as expected");

cleanup:

	xt_file_remove(xth);
	xt_free(xth);
}

void
test_setxattr_create()

{
	xt_hdl *xth;
	int ret;
	ssize_t xlen;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	ret = xt_file_create(xth);
	if (ret < 0) {
		fail(xth, "file create errno %d (%s)", errno,
		    strerror(errno));
		return;
	}

	ret = xt_value_alloc(xth, 16);
	if (ret < 0) {
		fail(xth, "value alloc errno %d (%s)", errno,
		    strerror(errno));
		goto cleanup;
	}

	/*
 	 * XATTR_CREATE for a non-existing xattr ==> OK
 	 */
	ret = xt_setxattr(xth, XATTR_CREATE);
	if (ret < 0) {
		fail(xth, "setxattr errno %d (%s)", errno, strerror(errno));
		goto cleanup;
	}

	/*
 	 * XATTR_CREATE for an existing xattr ==> EEXIST
 	 */
	ret = xt_setxattr(xth, XATTR_CREATE);
	if (ret != -EEXIST) {
		fail(xth, "expected %d got %d (%s)", EEXIST,
		    -ret, ret == 0 ? "no error" : strerror(-ret));
		goto cleanup;
	}

	/*
	 * Verify value (xattr should still exist).
	 */
	xt_value_clear(xth);

	xlen = xt_getxattr(xth);
	if (xlen < 0) {
		fail(xth, "getxattr errno %d (%s)", errno, strerror(errno));
		goto cleanup;
	}

	if (xlen != xt_value_len(xth)) {
		fail(xth, "getxattr length mismatch (expected %lld got %lld)",
		    (long long)xt_value_len(xth), (long long)xlen);
		goto cleanup;
	}

	ret = xt_value_check(xth, 16);
	if (ret < 0) {
		fail(xth, "value check failed at offset %lld",
		    xt_value_failoff(xth));
		goto cleanup;
	}

	pass(xth, "setxattr with XATTR_CREATE behaves correctly");

cleanup:

	xt_file_remove(xth);
	xt_free(xth);
}

void
test_setxattr_replace()

{
	xt_hdl *xth;
	int ret;
	ssize_t xlen;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	ret = xt_file_create(xth);
	if (ret < 0) {
		fail(xth, "file create errno %d (%s)", errno,
		    strerror(errno));
		return;
	}

	ret = xt_value_alloc(xth, 16);
	if (ret < 0) {
		fail(xth, "value alloc errno %d (%s)", errno,
		    strerror(errno));
		goto cleanup;
	}

	/*
 	 * XATTR_REPLACE for a non-existing xattr ==> ENODATA
 	 */
	ret = xt_setxattr(xth, XATTR_REPLACE);
	if (ret != -ENODATA) {
		fail(xth, "expected %d got %d (%s)", ENODATA,
		    -ret, ret == 0 ? "no error" : strerror(-ret));
		goto cleanup;
	}

	/*
 	 * Create it, try to replace it afterwards.
 	 */
	ret = xt_setxattr(xth, 0);
	if (ret < 0) {
		fail(xth, "setxattr errno %d (%s)", errno, strerror(errno));
		goto cleanup;
	}

	xt_value_fill(xth);

	/*
 	 * XATTR_REPLACE should work.
 	 */
	ret = xt_setxattr(xth, XATTR_REPLACE);
	if (ret < 0) {
		fail(xth, "setxattr errno %d (%s)", errno, strerror(errno));
		goto cleanup;
	}

	/*
	 * Verify value (xattr should still exist).
	 */
	xt_value_clear(xth);

	xlen = xt_getxattr(xth);
	if (xlen < 0) {
		fail(xth, "getxattr errno %d (%s)", errno, strerror(errno));
		goto cleanup;
	}

	if (xlen != xt_value_len(xth)) {
		fail(xth, "getxattr length mismatch (expected %lld got %lld)",
		    (long long)xt_value_len(xth), (long long)xlen);
		goto cleanup;
	}

	ret = xt_value_check(xth, 16);
	if (ret < 0) {
		fail(xth, "value check failed at offset %lld",
		    xt_value_failoff(xth));
		goto cleanup;
	}

	pass(xth, "setxattr with XATTR_REPLACE behaves correctly");

cleanup:

	xt_file_remove(xth);
	xt_free(xth);
}

void
test_setxattr_badflag()

{
	xt_hdl *xth;
	int ret;
	ssize_t xlen;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	ret = xt_file_create(xth);
	if (ret < 0) {
		fail(xth, "file create errno %d (%s)", errno,
		    strerror(errno));
		return;
	}

	ret = xt_value_alloc(xth, 16);
	if (ret < 0) {
		fail(xth, "value alloc errno %d (%s)", errno,
		    strerror(errno));
		goto cleanup;
	}

	/*
 	 * XATTR_REPLACE for a non-existing xattr ==> ENODATA
 	 */
	ret = xt_setxattr(xth, XATTR_REPLACE);
	if (ret != -ENODATA) {
		fail(xth, "expected %d got %d (%s)", ENODATA,
		    -ret, ret == 0 ? "no error" : strerror(-ret));
		goto cleanup;
	}

	ret = xt_setxattr(xth, -1);
	if (ret != -EINVAL) {
		fail(xth, "setxattr errno %d (%s)", errno, strerror(errno));
		goto cleanup;
	}

	pass(xth, "setxattr with a bad flag behaves correctly");

cleanup:

	xt_file_remove(xth);
	xt_free(xth);
}

void
test_removexattr()

{
	xt_hdl *xth;
	int ret;
	ssize_t xlen;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	ret = xt_file_create(xth);
	if (ret < 0) {
		fail(xth, "file create errno %d (%s)", errno,
		    strerror(errno));
		return;
	}

	ret = xt_value_alloc(xth, 16);
	if (ret < 0) {
		fail(xth, "value alloc errno %d (%s)", errno,
		    strerror(errno));
		goto cleanup;
	}

	ret = xt_removexattr(xth);
	if (ret != -ENODATA) {
		fail(xth, "expected %d got %d (%s)", ENODATA,
		    -ret, ret == 0 ? "no error" : strerror(-ret));
		goto cleanup;
	}

	/*
 	 * Create it, try to remove it afterwards.
 	 */
	ret = xt_setxattr(xth, 0);
	if (ret < 0) {
		fail(xth, "setxattr errno %d (%s)", errno, strerror(errno));
		goto cleanup;
	}

	ret = xt_removexattr(xth);
	if (ret < 0) {
		fail(xth, "removexattr errno %d (%s)", errno,
		    strerror(errno));
		goto cleanup;
	}

	/*
	 * See if it's actually gone.
	 */
	xlen = xt_getxattr(xth);
	if (xlen != -ENODATA) {
		fail(xth, "expected %d got %d (%s)", ENODATA,
		    -xlen, xlen == 0 ? "no error" : strerror(xlen));
		goto cleanup;
	}

	pass(xth, "removexattr behaves correctly");

cleanup:

	xt_file_remove(xth);
	xt_free(xth);
}

void
__test_listxattr(xt_hdl *xth, ssize_t nattr)

{
	ssize_t len, alen;
	char **list1, **list2;
	int ret;

	list1 = list2 = NULL;

	ret = xattrs_alloc(&list1, nattr, &alen);
	if (ret < 0) {
		fail(xth, "failed to allocate xattr names");
		return;
	}

	ret = xattrs_listset(xt_filename(xth), list1, nattr);
	if (ret < 0) {
		fail(xth, "failed to set xattrs");
		goto out;
	}

	ret = xattrs_file2list(xt_filename(xth), &list2, &len);
	if (ret < 0) {
		if (alen > XATTR_SIZE_MAX && ret == -E2BIG) {
			/*
			 * Expected failure.
			 */
			goto rm;
		}
		fail(xth, "failed to get xattrs");
		goto out;
	}

	if (xattrs_cmplists(list1, nattr, list2, len)) {
		ret = -EINVAL;	/* XXX */
		fail(xth, "list comparison failed");
		goto out;
		
	}

rm:

	ret = xattrs_listrm(xt_filename(xth), list1, nattr);
	if (ret < 0) {
		fail(xth, "failed to remove xattrs");
		goto out;
	}

	pass(xth, "expected results listing and removing %lld xattrs",
	    (long long)nattr);

out:

	if (list2 != NULL)
		xattrs_free(list2, len);
	if (list1 != NULL)
		xattrs_free(list1, nattr);
}

void
test_listxattr_1()

{
	xt_hdl *xth;
	int ret;
	char **list1, **list2;
	ssize_t len;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	ret = xt_file_create(xth);
	if (ret < 0) {
		fail(xth, "file create errno %d (%s)", errno,
		    strerror(errno));
		return;
	}

	__test_listxattr(xth, 1);

	xt_file_remove(xth);
	xt_free(xth);
}

void
test_listxattr_256()

{
	xt_hdl *xth;
	int ret;
	char **list1, **list2;
	ssize_t len;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	ret = xt_file_create(xth);
	if (ret < 0) {
		fail(xth, "file create errno %d (%s)", errno,
		    strerror(errno));
		return;
	}

	__test_listxattr(xth, 256);

	xt_file_remove(xth);
	xt_free(xth);
}

void
test_listxattr_large()

{
	xt_hdl *xth;
	int ret;
	char **list1, **list2;
	ssize_t len;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	ret = xt_file_create(xth);
	if (ret < 0) {
		fail(xth, "file create errno %d (%s)", errno,
		    strerror(errno));
		return;
	}

	__test_listxattr(xth, XATTR_TEST_NATTR_MAX - 2);

	xt_file_remove(xth);
	xt_free(xth);
}

void
test_listxattr_2big()

{
	xt_hdl *xth;
	int ret;
	char **list1, **list2;
	ssize_t len;

	xth = xt_init(__func__);
	if (xth == NULL) {
		printf("%-24s FAIL: can't initialize handle", __func__);
		return;
	}

	ret = xt_file_create(xth);
	if (ret < 0) {
		fail(xth, "file create errno %d (%s)", errno,
		    strerror(errno));
		return;
	}

	__test_listxattr(xth, XATTR_TEST_NATTR_MAX + 1);

	xt_file_remove(xth);
	xt_free(xth);
}

int
main(int argc, char **argv)

{
	struct stat st;
	int fd, ret;

	if (argc < 2)
		usage();

	ret = xt_set_fs(argv[1]);
	if (ret != 0) {
		fprintf(stderr, "xattr: set fs: %s\n", strerror(ret));
		exit(1);
	}

	srandom(time(NULL));

	test_xattr_probe();
	test_xattr_0();
	test_xattr_1();
	test_xattr_256();
	test_xattr_1page();
	test_xattr_2pages();
	test_xattr_max();
	test_xattr_toolarge();

	test_setxattr_create();
	test_setxattr_replace();
	test_setxattr_badflag();

	test_removexattr();

	test_listxattr_1();
	test_listxattr_256();
	test_listxattr_large();
	test_listxattr_2big();

	return 0;
}
#24 lore raw
In reply to Frank van der Linden

By the way, I can't remember if I asked this before: is there a particular use case that motivates this xattr work?

--b.

Quoted 69 lines · Frank van der Linden

v3:
* Rebase to v5.8-rc2
* Use length probe + allocate + query for the listxattr and setxattr operations to avoid allocating unneeded space.
* Because of the above, drop the 'use kvmalloc for svcxdr_tmpalloc' patch, as it's no longer needed.

v2:
* As per the discussion, user extended attributes are enabled if the client and server support them (e.g. they support 4.2 and advertise the user extended attribute FATTR). There are no longer options to switch them off.
* The code is no longer conditioned on a config option.
* The number of patches has been reduced somewhat by merging smaller, related ones.
* Renamed some functions and added parameter comments as requested.

v1:

* Split in to client and server (changed from the original RFC patch).

Original RFC combined set is here:

https://www.spinics.net/lists/linux-nfs/msg74843.html

In general, these patches were, both server and client, tested as follows:
* stress-ng-xattr with 1000 workers
* Test all corner cases (XATTR_SIZE_*)
* Test all failure cases (no xattr, setxattr with different or invalid flags, etc).
* Verify the content of xattrs across several operations.
* Use KASAN and KMEMLEAK for a longer mix of testruns to verify that there were no leaks (after unmounting the filesystem).

(quoted diff) +0 -0
 	* Interop run against FreeBSD server/client implementation.
 	* Ran xfstests-dev, with no unexpected/new failures as compared

to an unpatched kernel. To fully use xfstests-dev, it needed some modifications, as it expects to either use all xattr namespaces, or none. Whereas NFS only suppors the "user." namespace (+ optional ACLs). I will send the changes in seperately.

Frank van der Linden (10):
xattr: break delegations in {set,remove}xattr
xattr: add a function to check if a namespace is supported nfs,nfsd: NFSv4.2 extended attribute protocol definitions
nfsd: split off the write decode code in to a separate function
nfsd: add defines for NFSv4.2 extended attribute support
nfsd: define xattr functions to call in to their vfs counterparts
nfsd: take xattr bits in to account for permission checks
nfsd: add structure definitions for xattr requests / responses
nfsd: implement the xattr functions and en/decode logic
nfsd: add fattr support for user extended attributes

 fs/nfsd/nfs4proc.c        | 128 ++++++++-
 fs/nfsd/nfs4xdr.c         | 531 +++++++++++++++++++++++++++++++++++---
 fs/nfsd/nfsd.h            |   5 +-
 fs/nfsd/vfs.c             | 239 +++++++++++++++++
 fs/nfsd/vfs.h             |  10 +
 fs/nfsd/xdr4.h            |  31 +++
 fs/xattr.c                | 111 +++++++-
 include/linux/nfs4.h      |  22 +-
 include/linux/xattr.h     |   4 +
 include/uapi/linux/nfs4.h |   3 +
 10 files changed, 1044 insertions(+), 40 deletions(-)
Signature · 1 line
2.17.2
#25 lore raw
Quoted 8 lines · J. Bruce Fields

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you can confirm the sender and know the content is safe.

By the way, I can't remember if I asked this before: is there a particular use case that motivates this xattr work?

--b.

There's one use case that I can't really talk about publicly at this point (and it's not my code either, so I wouldn't have all the details). Nothing super secret or anything - it's just something that is not mine, so I won't try to speak for anyone. We wanted to get this upstreamed first, as that's the right thing to do.

Since I posted my first RFC, I did get contacted off-list by several readers of linux-nfs who wanted to use the feature in practice, too, so there's definitely interest out there.

- Frank

#26 lore raw
Quoted 12 lines · Frank van der Linden
Quoted 4 lines · J. Bruce Fields

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you can confirm the sender and know the content is safe.

By the way, I can't remember if I asked this before: is there a particular use case that motivates this xattr work?

There's one use case that I can't really talk about publicly at this point (and it's not my code either, so I wouldn't have all the details). Nothing super secret or anything - it's just something that is not mine, so I won't try to speak for anyone. We wanted to get this upstreamed first, as that's the right thing to do.

Quoted 3 lines

Since I posted my first RFC, I did get contacted off-list by several readers of linux-nfs who wanted to use the feature in practice, too, so there's definitely interest out there.

Yeah, I always hear a lot of interest but then have trouble sorting through it for the cases that are actually *user* xattr cases, where the server has to just act as a dumb store of the values.

There are some. But unfortunately xattrs are best known for enabling selinux and posix acls, and to a lesser extent accessing random other filesystem features, so that tends to be what comes to people's minds first, though it's not what we're doing.

--b.

#27 lore raw
In reply to Frank van der Linden
Quoted 10 lines · Frank van der Linden

v3:
* Rebase to v5.8-rc2
* Use length probe + allocate + query for the listxattr and setxattr operations to avoid allocating unneeded space.
* Because of the above, drop the 'use kvmalloc for svcxdr_tmpalloc' patch, as it's no longer needed.

v3 of this series has been applied to nfsd-5.9. Thanks!

See: git://git.linux-nfs.org/projects/cel/cel-2.6.git nfsd-5.9

Still waiting for Acks on 01/13 and 02/13.

Quoted 48 lines

v2:
* As per the discussion, user extended attributes are enabled if the client and server support them (e.g. they support 4.2 and advertise the user extended attribute FATTR). There are no longer options to switch them off.
* The code is no longer conditioned on a config option.
* The number of patches has been reduced somewhat by merging smaller, related ones.
* Renamed some functions and added parameter comments as requested.

v1:

* Split in to client and server (changed from the original RFC patch).

Original RFC combined set is here:

https://www.spinics.net/lists/linux-nfs/msg74843.html

In general, these patches were, both server and client, tested as follows:
* stress-ng-xattr with 1000 workers
* Test all corner cases (XATTR_SIZE_*)
* Test all failure cases (no xattr, setxattr with different or invalid flags, etc).
* Verify the content of xattrs across several operations.
* Use KASAN and KMEMLEAK for a longer mix of testruns to verify that there were no leaks (after unmounting the filesystem).
* Interop run against FreeBSD server/client implementation.
* Ran xfstests-dev, with no unexpected/new failures as compared to an unpatched kernel. To fully use xfstests-dev, it needed some modifications, as it expects to either use all xattr namespaces, or none. Whereas NFS only suppors the "user." namespace (+ optional ACLs). I will send the changes in seperately.

Frank van der Linden (10):
xattr: break delegations in {set,remove}xattr
xattr: add a function to check if a namespace is supported nfs,nfsd: NFSv4.2 extended attribute protocol definitions
nfsd: split off the write decode code in to a separate function
nfsd: add defines for NFSv4.2 extended attribute support
nfsd: define xattr functions to call in to their vfs counterparts
nfsd: take xattr bits in to account for permission checks
nfsd: add structure definitions for xattr requests / responses
nfsd: implement the xattr functions and en/decode logic
nfsd: add fattr support for user extended attributes

Quoted 15 lines
fs/nfsd/nfs4proc.c        | 128 ++++++++-
fs/nfsd/nfs4xdr.c         | 531 +++++++++++++++++++++++++++++++++++---
fs/nfsd/nfsd.h            |   5 +-
fs/nfsd/vfs.c             | 239 +++++++++++++++++
fs/nfsd/vfs.h             |  10 +
fs/nfsd/xdr4.h            |  31 +++
fs/xattr.c                | 111 +++++++-
include/linux/nfs4.h      |  22 +-
include/linux/xattr.h     |   4 +
include/uapi/linux/nfs4.h |   3 +
10 files changed, 1044 insertions(+), 40 deletions(-)
Signature · 1 line
2.17.2

--
Chuck Lever